Go to main content
Oracle Developer Studio 12.5 Man Pages

Exit Print View

Updated: June 2017
 
 

libsunperf (3p)

Name

libsunperf - formance Library functions and subroutines

Synopsis

Please see following description for synopsis

Description

Oracle Solaris Studio Performance Library                            intro(3P)



NAME
       intro: sunperf, libsunperf - Introduction to Oracle Solaris Studio Per-
       formance Library functions and subroutines


DESCRIPTION
       Oracle Solaris Studio Performance Library (sunperf) is a set  of  opti-
       mized,  high-speed  mathematical subroutines for solving linear algebra
       and other numerically intensive problems. Oracle Solaris Studio Perfor-
       mance  Library  is  based on a collection of public domain applications
       available from Netlib at http://www.netlib.org. Oracle  Solaris  Studio
       has  enhanced  these public domain applications and bundled them as the
       Oracle Solaris Studio Performance Library.

       More information about Oracle Solaris Studio Performance Library can be
       found in the Oracle Solaris Studio Performance Library User's Guide and
       Oracle Solaris Studio Performance Library Reference Manual.


LIBRARIES
       Oracle Solaris Studio Performance Library contains enhanced versions of
       the following standard libraries:



       Library            Version  Description

       LAPACK             3.1.1    solving linear algebra problems
       BLAS1                -      performing vector-vector operations
       BLAS2                -      performing matrix-vector operations
       BLAS3                -      performing matrix-matrix operations
       Netlib Sparse-BLAS   -      performing sparse vector operations
       NIST Sparse-BLAS    0.5     performing fundamental sparse matrix operations
       SuperLU             3.0     solving sparse linear systems of equations
       Sparse Solver        -      direct sparse solver routines
       FFTPACK              -      performing fast Fourier transform
       VFFTPACK             -      performing vectorized fast Fourier transform
       IBLAS                -      interval BLAS routines
       Other Routines       -      trabspose, Convolution, correlation and sort

       A list of the individual subroutines is included at the bottom of this page.


FEATURES
       Oracle  Solaris Studio Performance Library routines can increase appli-
       cation performance on both serial and  multiprocessor  (MP)  platforms,
       because  the  serial  speed  of  many Oracle Solaris Studio Performance
       Library routines has been increased, and many routines have been paral-
       lelized.  Oracle  Solaris Studio Performance Library routines also have
       SPARC, AMD and Intel specific optimizations that are not present in the
       base Netlib libraries.

       Oracle  Solaris Studio Performance Library provides the following opti-
       mizations and extensions to the base Netlib libraries:


                     Extensions that support Fortran 95 and C language  inter-
                 faces


                     Fortran  95  language  features,  including type indepen-
                 dence, compile time checking, and optional arguments.


                     Consistent API across the different libraries  in  Oracle
                 Solaris Studio Performance Library


                     Compatibility   with   LAPACK  1,  2.0,  3.0,  and  3.1.1
                 libraries


                     Increased performance, and in some cases,  greater  accu-
                 racy


                     Optimizations  for specific SPARC and x86/x64 instruction
                 set architectures


                     Support for 64-bit enabled Solaris  and  Linux  operating
                 environments


                     Support  for  parallel  processing  compiler  options for
                 SPARC and x86/x64 platforms


                     Support for multiple processor hardware options

USAGE
       To use the Oracle Solaris Studio Performance Library, type one  of  the
       following commands.

       % f95 -dalign file.f -library=sunperf

       or

       % cc -dalign file.c -library=sunperf

       or

       % CC -dalign file.c -library=sunperf

       To  link with the Oracle Solaris Studio Performance Library statically,
       add -staticlib=sunperf to the commandline.


SUBROUTINES
       Copy_CompCol_Matrix - A utility C function in the serial SuperLU solver
                 that copies one SuperMatrix into another.

       Create_CompCol_Matrix  -  A  utility  C  function in the serial SuperLU
                 solver that creates a SuperMatrix in compressed sparse column
                 format (also known as the Harwell-Boeing format).

       Create_CompRow_Matrix  -  A  utility  C  function in the serial SuperLU
                 solver that creates a SuperMatrix in  compressed  sparse  row
                 format.

       Create_Dense_Matrix - A utility C function in the serial SuperLU solver
                 that creates a SuperMatrix in dense format.

       Create_SuperNode_Matrix - A utility C function in  the  serial  SuperLU
                 solver that creates a SuperMatrix in supernodal format.

       Destroy_CompCol_Matrix - Precision-independent C function in the serial
                 SuperLU solver that deallocates a supermatrix  in  compressed
                 sparse  column  format (also known as the Harwell-Boeing for-
                 mat).

       Destroy_CompCol_Permuted -  Precision-independent  C  function  in  the
                 serial  SuperLU solver that deallocates a supermatrix in per-
                 muted, compressed sparse column format.

       Destroy_CompRow_Matrix - Precision-independent C function in the serial
                 SuperLU  solver  that deallocates a supermatrix in compressed
                 sparse row format.

       Destroy_Dense_Matrix - Precision-independent C function in  the  serial
                 SuperLU  solver  that deallocates a SuperMatrix in dense for-
                 mat.

       Destroy_SuperMatrix_Store - Precision-independent  C  function  in  the
                 serial  SuperLU  solver  that  deallocates the actual storage
                 used to store the matrix in a SuperMatrix.

       Destroy_SuperNode_Matrix -  Precision-independent  C  function  in  the
                 serial  SuperLU  solver  that  deallocates  a  SuperMatrix in
                 supernodal format.

       LUFactFlops - A query function that returns the floating  point  opera-
                 tion count of the factorization step of the SuperLU solver.

       LUFactTime  -  A query function that returns the time spent in the fac-
                 torization step by the SuperLU solver.

       LUSolveFlops - A query function that returns the floating point  opera-
                 tion count of the solve step of the SuperLU solver.

       LUSolveTime - A query function that returns the time spent in the solve
                 stage by the SuperLU solver.

       PrintPerf - A utility function of the SuperLU solver that  prints  sta-
                 tistics collected by the computational routines.

       QuerySpace - A inquiry function that provides information on the memory
                 statistics of the SuperLU solver.

       StatFree - frees storage that was previously allocated to hold  perfor-
                 mance statistics of the SuperLU solver.

       StatInit  -  A  utility C function that allocates and initializes vari-
                 ables in structure that stores  performance  statistics  col-
                 lected during the computation of the SuperLU solver.

       SuperMatrix  - C data structure in the SuperLU software that represents
                 a sparse or dense general matrix.

       available_threads - returns information about current thread usage

       blas_dpermute - permutes a real (double precision) array  in  terms  of
                 the permutation vector P, output by dsortv

       blas_dsort  - sorts a real (double precision) vector X in increasing or
                 decreasing order using quick sort algorithm

       blas_dsortv - sorts a real (double precision) vector X in increasing or
                 decreasing  order  using quick sort algorithm and overwrite P
                 with the permutation vector

       blas_ipermute - permutes an integer array in terms of  the  permutation
                 vector P, output by dsortv

       blas_isort  -  sorts  an  integer  vector X in increasing or decreasing
                 order using quick sort algorithm

       blas_isortv - sorts a real vector X in increasing or  decreasing  order
                 using  quick sort algorithm and overwrite P with the permuta-
                 tion vector

       blas_spermute - permutes a real array in terms of the permutation  vec-
                 tor P, output by dsortv

       blas_ssort  -  sorts  a real vector X in increasing or decreasing order
                 using quick sort algorithm

       blas_ssortv - sorts a real vector X in increasing or  decreasing  order
                 using  quick sort algorithm and overwrite P with the permuta-
                 tion vector

       cCopy_CompCol_Matrix - A utility  C  function  in  the  serial  SuperLU
                 solver that copies one SuperMatrix into another.

       cCreate_CompCol_Matrix  -  A  utility  C function in the serial SuperLU
                 solver that creates a SuperMatrix in compressed sparse column
                 format (also known as the Harwell-Boeing format).

       cCreate_CompRow_Matrix  -  A  utility  C function in the serial SuperLU
                 solver that creates a SuperMatrix in  compressed  sparse  row
                 format.

       cCreate_Dense_Matrix  -  A  utility  C  function  in the serial SuperLU
                 solver that creates a SuperMatrix in dense format.

       cCreate_SuperNode_Matrix - A utility C function in the  serial  SuperLU
                 solver that creates a SuperMatrix in supernodal format.

       cPrintPerf  - A utility function of the SuperLU solver that prints sta-
                 tistics collected by the computational routines.

       cQuerySpace - A inquiry function that provides information on the  mem-
                 ory statistics of the SuperLU solver.

       caxpy - compute y := alpha * x + y

       caxpyi - Compute y := alpha * x + y

       cbbcsd  -  compute the CS decomposition of a unitary matrix in bidiago-
                 nal-block form

       cbcomm - block coordinate matrix-matrix multiply

       cbdimm - block diagonal format matrix-matrix multiply

       cbdism -  block diagonal format triangular solve

       cbdsqr - compute the singular value decomposition (SVD) of a real N-by-
                 N (upper or lower) bidiagonal matrix B

       cbelmm - block Ellpack format matrix-matrix multiply

       cbelsm - block Ellpack format triangular solve

       cblas  -  C  interface  to  the  original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_caxpy - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ccopy - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_cdotc_sub - C interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_cdotu_sub - C interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_cgbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_cgemm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_cgemv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_cgerc - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_cgeru - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_chbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_chemm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_chemv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_cher - C interface to the original Level 1, 2 and 3 BLAS, or  the
                 Legacy BLAS

       cblas_cher2 - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_cher2k - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_cherk - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_chpmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_chpr  - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_chpr2 - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_cscal - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_csscal - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_cswap - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_csymm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_csyr2k  -  C  interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_csyrk - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ctbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ctbsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ctpmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ctpsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ctrmm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ctrmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ctrsm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ctrsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dasum - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_daxpy - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dcopy - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ddot - C interface to the original Level 1, 2 and 3 BLAS, or  the
                 Legacy BLAS

       cblas_dgbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dgemm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dgemv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dger - C interface to the original Level 1, 2 and 3 BLAS, or  the
                 Legacy BLAS

       cblas_dnrm2 - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_drot - C interface to the original Level 1, 2 and 3 BLAS, or  the
                 Legacy BLAS

       cblas_drotg - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_drotm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_drotmg  -  C  interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_dsbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dscal - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dsdot - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dspmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dspr - C interface to the original Level 1, 2 and 3 BLAS, or  the
                 Legacy BLAS

       cblas_dspr2 - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dswap - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dsymm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dsymv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dsyr  - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dsyr2 - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dsyr2k  -  C  interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_dsyrk - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dtbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dtbsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dtpmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dtpsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dtrmm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dtrmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dtrsm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dtrsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_dzasum  -  C  interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_dznrm2 - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_icamax  -  C  interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_idamax - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_isamax  -  C  interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_izamax - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_memerr  -  C  interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_sasum - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_saxpy - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_scasum - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_scnrm2  -  C  interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_scopy - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_sdot  - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_sdsdot - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_sgbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_sgemm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_sgemv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_sger - C interface to the original Level 1, 2 and 3 BLAS, or  the
                 Legacy BLAS

       cblas_snrm2 - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_srot - C interface to the original Level 1, 2 and 3 BLAS, or  the
                 Legacy BLAS

       cblas_srotg - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_srotm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_srotmg  -  C  interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_ssbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_sscal - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_sspmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_sspr  - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_sspr2 - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_sswap - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ssymm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ssymv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ssyr - C interface to the original Level 1, 2 and 3 BLAS, or  the
                 Legacy BLAS

       cblas_ssyr2 - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ssyr2k - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_ssyrk - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_stbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_stbsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_stpmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_stpsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_strmm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_strmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_strsm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_strsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_xerbla - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_zaxpy - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zcopy - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zdotc_sub - C interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_zdotu_sub - C interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_zdscal  -  C  interface to the original Level 1, 2 and 3 BLAS, or
                 the Legacy BLAS

       cblas_zgbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zgemm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zgemv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zgerc - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zgeru - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zhbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zhemm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zhemv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zher - C interface to the original Level 1, 2 and 3 BLAS, or  the
                 Legacy BLAS

       cblas_zher2 - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zher2k - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_zherk - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zhpmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zhpr  - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zhpr2 - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zscal - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zswap - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zsymm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_zsyr2k - C interface to the original Level 1, 2 and  3  BLAS,  or
                 the Legacy BLAS

       cblas_zsyrk - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ztbmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ztbsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ztpmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ztpsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ztrmm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ztrmv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ztrsm - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cblas_ztrsv - C interface to the original Level 1, 2 and 3 BLAS, or the
                 Legacy BLAS

       cbscmm - block sparse column matrix-matrix multiply

       cbscsm - block sparse column format triangular solve

       cbsrmm - block sparse row format matrix-matrix multiply

       cbsrsm - block sparse row format triangular solve

       ccnvcor - compute the convolution or correlation of complex vectors

       ccnvcor2 - compute the convolution or correlation of complex matrices

       ccoomm - coordinate matrix-matrix multiply

       ccopy - copy x to y

       ccscmm - compressed sparse column format matrix-matrix multiply

       ccscsm - compressed sparse column format triangular solve

       ccsrmm - compressed sparse row format matrix-matrix multiply

       ccsrsm - compressed sparse row format triangular solve

       cdiamm - diagonal format matrix-matrix multiply

       cdiasm - diagonal format triangular solve

       cdotc - compute the dot product of two vectors conjg(x) and y.

       cdotci - Compute the complex conjugated indexed dot product.

       cdotu - compute the dot product of two vectors x and y.

       cdotui - Compute the complex unconjugated indexed dot product.

       cellmm - Ellpack format matrix-matrix multiply

       cellsm - Ellpack format triangular solve

       cfft2b - compute a periodic sequence  from  its  Fourier  coefficients.
                 The  xFFT  operations  are  unnormalized, so a call of xFFT2F
                 followed by a call of xFFT2B will multiply the input sequence
                 by M*N.

       cfft2f  - compute the Fourier coefficients of a periodic sequence.  The
                 xFFT operations are unnormalized, so a call  of  xFFT2F  fol-
                 lowed by a call of xFFT2B will multiply the input sequence by
                 M*N.

       cfft2i - initialize the array WSAVE, which is used in both the  forward
                 and backward transforms.

       cfft3b  -  compute  a  periodic sequence from its Fourier coefficients.
                 The FFT operations are unnormalized, so a call of CFFT3F fol-
                 lowed by a call of CFFT3B will multiply the input sequence by
                 M*N*K.

       cfft3f - compute the Fourier coefficients of a periodic sequence.   The
                 FFT operations are unnormalized, so a call of CFFT3F followed
                 by a call of CFFT3B  will  multiply  the  input  sequence  by
                 M*N*K.

       cfft3i  -  initialize the array WSAVE, which is used in both CFFT3F and
                 CFFT3B.

       cfftb - compute a periodic sequence from its Fourier coefficients.  The
                 FFT  operations are unnormalized, so a call of CFFTF followed
                 by a call of CFFTB will multiply the input sequence by N.

       cfftc - initialize the trigonometric weight and factor tables  or  com-
                 pute  the  Fast  Fourier  transform (forward or inverse) of a
                 complex sequence.

       cfftc2 - initialize the trigonometric weight and factor tables or  com-
                 pute  the  two-dimensional Fast Fourier Transform (forward or
                 inverse) of a two-dimensional complex array.

       cfftc3 - initialize the trigonometric weight and factor tables or  com-
                 pute the three-dimensional Fast Fourier Transform (forward or
                 inverse) of a three-dimensional complex array.

       cfftcm - initialize the trigonometric weight and factor tables or  com-
                 pute  the  one-dimensional Fast Fourier Transform (forward or
                 inverse) of a set of data sequences stored  in  a  two-dimen-
                 sional complex array.

       cfftf  -  compute the Fourier coefficients of a periodic sequence.  The
                 FFT operations are unnormalized, so a call of CFFTF  followed
                 by a call of CFFTB will multiply the input sequence by N.

       cffti  -  initialize  the  array WSAVE, which is used in both CFFTF and
                 CFFTB.

       cfftopt - compute the length of the closest fast FFT

       cffts - initialize the trigonometric weight and factor tables  or  com-
                 pute the inverse Fast Fourier Transform of a complex sequence
                 as follows.

       cffts2 - initialize the trigonometric weight and factor tables or  com-
                 pute  the two-dimensional inverse Fast Fourier Transform of a
                 two-dimensional complex array.

       cffts3 - initialize the trigonometric weight and factor tables or  com-
                 pute  the three-dimensional inverse Fast Fourier Transform of
                 a three-dimensional complex array.

       cfftsm - initialize the trigonometric weight and factor tables or  com-
                 pute  the one-dimensional inverse Fast Fourier Transform of a
                 set of complex data sequences  stored  in  a  two-dimensional
                 array.

       cgbbrd  -  reduce  a complex general m-by-n band matrix A to real upper
                 bidiagonal form B by a unitary transformation

       cgbcon - estimate the reciprocal of the condition number of  a  complex
                 general  band matrix A, in either the 1-norm or the infinity-
                 norm, using the LU factorization computed by CGBTRF

       cgbequ - compute row and column scalings intended to equilibrate an  M-
                 by-N band matrix A and reduce its condition number

       cgbequb - compute row and column scalings intended to equilibrate an M-
                 by-N matrix A and reduce its condition number

       cgbmv - perform one of  the  matrix-vector  operations  y:=alpha*A*x  +
                 beta*y,  or y:=alpha*A'*x + beta*y, or y:=alpha*conjg(A')*x +
                 beta*y

       cgbrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix  is  banded, and provide error
                 bounds and backward error estimates for the solution

       cgbrfsx - improve the computed solution to a system of linear equations
                 and provide error bounds and backward error estimates for the
                 solution

       cgbsv - compute the solution to a complex system  of  linear  equations
                 A*X=B,  where A is a band matrix of order N with KL subdiago-
                 nals and KU superdiagonals, and X and B are N-by-NRHS  matri-
                 ces

       cgbsvx  - use the LU factorization to compute the solution to a complex
                 system of linear  equations  A*X=B,  A**T*X=B,  or  A**H*X=B,
                 where A is a band matrix

       cgbsvxx  - compute the solution to system of linear equations A * X = B
                 for ganeral band matrices

       cgbtf2 - compute an LU factorization of a complex m-by-n band matrix  A
                 using partial pivoting with row interchanges

       cgbtrf  - compute an LU factorization of a complex m-by-n band matrix A
                 using partial pivoting with row interchanges

       cgbtrs - solve  a  system  of  linear  equations  A*X=B,  A**T*X=B,  or
                 A**H*X=B with a general band matrix A using the LU factoriza-
                 tion computed by CGBTRF

       cgebak - form the right or  left  eigenvectors  of  a  complex  general
                 matrix  by  backward transformation on the computed eigenvec-
                 tors of the balanced matrix output by CGEBAL

       cgebal - balance a general complex matrix A

       cgebd2 - reduce a general matrix to bidiagonal form using an  unblocked
                 algorithm

       cgebrd  -  reduce  a  general complex M-by-N matrix A to upper or lower
                 bidiagonal form B by a unitary transformation

       cgecon - estimate the reciprocal of the condition number of  a  general
                 complex  matrix A, in either the 1-norm or the infinity-norm,
                 using the LU factorization computed by CGETRF

       cgeequ - compute row and column scalings intended to equilibrate an  M-
                 by-N matrix A and reduce its condition number

       cgeequb - compute row and column scalings intended to equilibrate an M-
                 by-N matrix A and reduce its condition number

       cgees - compute for an N-by-N complex nonsymmetric matrix A, the eigen-
                 values,  the  Schur  form  T,  and, optionally, the matrix of
                 Schur vectors Z

       cgeesx - compute for an N-by-N complex nonsymmetric matrix A,  the  ei-
                 genvalues,  the  Schur form T, and, optionally, the matrix of
                 Schur vectors Z

       cgeev - compute for an N-by-N complex nonsymmetric matrix A, the eigen-
                 values and, optionally, the left and/or right eigenvectors

       cgeevx  -  compute for an N-by-N complex nonsymmetric matrix A, the ei-
                 genvalues and, optionally, the left and/or right eigenvectors

       cgegs - routine is deprecated and has been replaced by routine CGGES

       cgegv - routine is deprecated and has been replaced by routine CGGEV

       cgehd2  - reduce a general square matrix to upper Hessenberg form using
                 an unblocked algorithm

       cgehrd - reduce a complex general matrix A to upper Hessenberg  form  H
                 by a unitary similarity transformation

       cgelq2  -  compute the LQ factorization of a general rectangular matrix
                 using an unblocked algorithm

       cgelqf - compute an LQ factorization of a complex M-by-N matrix A

       cgels - solve overdetermined or underdetermined complex linear  systems
                 involving  an  M-by-N  matrix  A, or its conjugate-transpose,
                 using a QR or LQ factorization of A

       cgelsd - compute the minimum-norm  solution  to  a  real  linear  least
                 squares problem

       cgelss  -  compute  the minimum norm solution to a complex linear least
                 squares problem

       cgelsx - routine is deprecated and has been replaced by routine CGELSY

       cgelsy - compute the minimum-norm solution to a  complex  linear  least
                 squares problem

       cgemm  -  perform  one of the matrix-matrix operations C := alpha*op( A
                 )*op( B ) + beta*C

       cgemqrt - overwrite the general complex M-by-N matrix C with Q*C,  C*Q,
                 Q**H*C, or C*Q**H depending on values of SIDE and TRANS

       cgemv  -  perform  one  of the matrix-vector operationsy := alpha*A*x +
                 beta*y, or y := alpha*A'*x + beta*y, or   y  :=  alpha*conjg(
                 A' )*x + beta*y

       cgeql2  -  compute the QL factorization of a general rectangular matrix
                 using an unblocked algorithm

       cgeqlf - compute a QL factorization of a complex M-by-N matrix A

       cgeqp3 - compute a QR factorization with column pivoting of a matrix A

       cgeqpf - routine is deprecated and has been replaced by routine CGEQP3

       cgeqr2 - computes the QR factorization of a general rectangular  matrix
                 using an unblocked algorithm.

       cgeqr2p - computes the QR factorization of a general rectangular matrix
                 with non-negative diagonal elements using an unblocked  algo-
                 rithm.

       cgeqrf - compute a QR factorization of a complex M-by-N matrix A

       cgeqrfp  - compute a QR factorization of a complex M-by-N matrix A: A =
                 Q * R

       cgeqrt - compute a blocked QR factorization of a complex M-by-N  matrix
                 A using the compact WY representation of Q

       cgeqrt2  - compute a QR factorization of a general complex matrix using
                 the compact WY representation of Q

       cgeqrt3 - recursively compute a QR factorization of a  general  complex
                 matrix using the compact WY representation of Q

       cgerc - perform the rank 1 operation A := alpha*x*conjg( y' ) + A

       cgerfs  - improve the computed solution to a system of linear equations
                 and provide error bounds and backward error estimates for the
                 solution

       cgerfsx - improve the computed solution to a system of linear equations
                 and provide error bounds and backward error estimates for the
                 solution

       cgerq2  - computes the RQ factorization of a general rectangular matrix
                 using an unblocked algorithm

       cgerqf - compute an RQ factorization of a complex M-by-N matrix A

       cgeru - perform the rank 1 operation A := alpha*x*y' + A

       cgesdd - compute the singular value decomposition (SVD) of a complex M-
                 by-N  matrix  A,  optionally  computing the left and/or right
                 singular vectors, by using divide-and-conquer method

       cgesv - compute the solution to a complex system  of  linear  equations
                 A*X=B,  where A is an N-by-N matrix and X and B are N-by-NRHS
                 matrices

       cgesvd - compute the singular value decomposition (SVD) of a complex M-
                 by-N  matrix  A,  optionally  computing the left and/or right
                 singular vectors

       cgesvx - use the LU factorization to compute the solution to a  complex
                 system  of  linear equations  A*X=B,  where is an N-by-N gen-
                 eral matrix

       cgesvxx - compute the solution to system of linear equations A*X=B  for
                 general matrices

       cgetf2 - compute an LU factorization of a general m-by-n matrix A using
                 partial pivoting with row interchanges

       cgetrf - compute an LU factorization of a general M-by-N matrix A using
                 partial pivoting with row interchanges

       cgetri  -  compute  the  inverse of a matrix using the LU factorization
                 computed by CGETRF

       cgetrs - solve a system of linear equations  A * X = B, A**T * X  =  B,
                 or  A**H  * X = B with a general N-by-N matrix A using the LU
                 factorization computed by CGETRF

       cggbak - form the right or left eigenvectors of a  complex  generalized
                 eigenvalue  problem A*x = lambda*B*x, by backward transforma-
                 tion on the computed eigenvectors of  the  balanced  pair  of
                 matrices output by CGGBAL

       cggbal - balance a pair of general complex matrices (A,B)

       cgges  -  compute  for  a  pair of N-by-N complex nonsymmetric matrices
                 (A,B), the generalized eigenvalues, the  generalized  complex
                 Schur  form  (S,  T),  and optionally left and/or right Schur
                 vectors (VSL and VSR)

       cggesx - compute for a pair of  N-by-N  complex  nonsymmetric  matrices
                 (A,B),  the  generalized  eigenvalues, the complex Schur form
                 (S,T), and, optionally, the left  and/or  right  matrices  of
                 Schur vectors

       cggev  -  compute  for  a  pair of N-by-N complex nonsymmetric matrices
                 (A,B), the generalized eigenvalues, and optionally, the  left
                 and/or right generalized eigenvectors

       cggevx  -  compute  for  a pair of N-by-N complex nonsymmetric matrices
                 (A,B) the generalized eigenvalues, and, optionally, the  left
                 and/or right generalized eigenvectors

       cggglm - solve a general Gauss-Markov linear model (GLM) problem

       cgghrd  -  reduce a pair of complex matrices (A,B) to generalized upper
                 Hessenberg form using unitary transformations, where A  is  a
                 general matrix and B is upper triangular

       cgglse  -  solve  the  linear  equality-constrained least squares (LSE)
                 problem

       cggqrf - compute a generalized QR factorization of an N-by-M  matrix  A
                 and an N-by-P matrix B.

       cggrqf  -  compute a generalized RQ factorization of an M-by-N matrix A
                 and a P-by-N matrix B

       cggsvd - compute the generalized singular value decomposition (GSVD) of
                 an M-by-N complex matrix A and P-by-N complex matrix B

       cggsvp - compute unitary matrices

       cgscon  - estimates the reciprocal of the condition number of a general
                 real matrix A, in either the  1-norm  or  the  infinity-norm,
                 using  the  LU  factorization  computed  by  SuperLU  routine
                 sgstrf.

       cgsequ - computes row and column scalings intended to equilibrate an M-
                 by-N sparse matrix A and reduce its condition number.

       cgsrfs - improves the computed solution to a system of linear equations
                 and provides error bounds and backward  error  estimates  for
                 the solution.  It is a SuperLU routine.

       cgssco - General sparse solver condition number estimate.

       cgssda - Deallocate working storage for the general sparse solver.

       cgssfa - General sparse solver numeric factorization.

       cgssfs - General sparse solver one call interface.

       cgssin - Initialize the general sparse solver.

       cgssor - General sparse solver ordering and symbolic factorization.

       cgssps - Print general sparse solver statics.

       cgssrp - Return permutation used by the general sparse solver.

       cgsssl - Solve routine for the general sparse solver.

       cgssuo  -  Provide  general sparse solvers SPSOLVE and SuperLU  a user-
                 supplied permutation for ordering.

       cgssv - solves a system of linear equations A*X=B using the LU  factor-
                 ization from sgstrf.

       cgssvx  -  solves the system of linear equations A*X=B or A'*X=B, using
                 the LU factorization from sgstrf(). Error bounds on the solu-
                 tion and a condition estimate are also provided.

       cgstrf - computes an LU factorization of a general sparse m-by-n matrix
                 A using partial pivoting with row interchanges.

       cgstrs - solves a system of linear equations A*X=B  or  A'*X=B  with  A
                 sparse  and  B  dense, using the LU factorization computed by
                 sgstrf.

       cgtcon - estimate the reciprocal of the condition number of  a  complex
                 tridiagonal  matrix  A using the LU factorization as computed
                 by CGTTRF

       cgthr - Gathers specified elements from y into x.

       cgthrz - Gather and zero.

       cgtrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix  is tridiagonal, provide error
                 bounds and backward error estimates for the solution

       cgtsv - solve the equation A*X = B, where A is  an  N-by-N  tridiagonal
                 matrix, by Gaussian elimination with partial pivoting

       cgtsvx  - use the LU factorization to compute the solution to a complex
                 system of linear  equations  A*X=B,  A**T*X=B,  or  A**H*X=B,
                 where A is a tridiagonal matrix of order N and X and B are N-
                 by-NRHS matrices

       cgttrf - compute an LU factorization of a complex tridiagonal matrix  A
                 using elimination with partial pivoting and row interchanges

       cgttrs  -  solve  one  of  the systems of equations A*X=B, A**T*X=B, or
                 A**H*X=B, with a tridiagonal matrix A using the LU factoriza-
                 tion computed by CGTTRF

       cgtts2  -  solve a system of linear equations with a tridiagonal matrix
                 using the LU factorization computed by cgttrf

       chbev - compute all the eigenvalues and, optionally, eigenvectors of  a
                 complex Hermitian band matrix A

       chbevd - compute all the eigenvalues and, optionally, eigenvectors of a
                 complex Hermitian band matrix A

       chbevx - compute selected eigenvalues and, optionally, eigenvectors  of
                 a complex Hermitian band matrix A

       chbgst  - reduce a complex Hermitian-definite banded generalized eigen-
                 problem A*x=lambda*B*x to standard  form  C*y=lambda*y,  such
                 that C has the same bandwidth as A

       chbgv  -  compute all the eigenvalues, and optionally, the eigenvectors
                 of a complex generalized Hermitian-definite banded eigenprob-
                 lem, of the form A*x=(lambda)*B*x

       chbgvd  - compute all the eigenvalues, and optionally, the eigenvectors
                 of a complex generalized Hermitian-definite banded eigenprob-
                 lem, of the form A*x=(lambda)*B*x

       chbgvx  - compute all the eigenvalues, and optionally, the eigenvectors
                 of a complex generalized Hermitian-definite banded eigenprob-
                 lem, of the form A*x=(lambda)*B*x

       chbmv - perform the matrix-vector operationy := alpha*A*x + beta*y

       chbtrd  -  reduce  a  complex Hermitian band matrix A to real symmetric
                 tridiagonal form T by a unitary similarity transformation

       checon - estimate the reciprocal of the condition number of  a  complex
                 Hermitian  matrix A using the factorization A = U*D*U**H or A
                 = L*D*L**H computed by CHETRF

       checon_rook - estimate the reciprocal of the condition number for  Her-
                 mitian  matrices using factorization obtained with one of the
                 bounded diagonal pivoting methods (max 2 interchanges)

       cheequb - compute row and column scalings  intended  to  equilibrate  a
                 Hermitian  matrix  A  and  reduce  its condition number (with
                 respect to the two-norm)

       cheev - compute all eigenvalues and, optionally, eigenvectors of a com-
                 plex Hermitian matrix A

       cheevd  -  compute  all  eigenvalues and, optionally, eigenvectors of a
                 complex Hermitian matrix A

       cheevr - compute selected eigenvalues and, optionally, eigenvectors  of
                 a complex Hermitian tridiagonal matrix T

       cheevx  - compute selected eigenvalues and, optionally, eigenvectors of
                 a complex Hermitian matrix A

       chegs2 - reduce a complex Hermitian-definite  generalized  eigenproblem
                 to standard form

       chegst  -  reduce a complex Hermitian-definite generalized eigenproblem
                 to standard form

       chegv - compute all the eigenvalues, and optionally,  the  eigenvectors
                 of  a complex generalized Hermitian-definite eigenproblem, of
                 the    form     A*x=(lambda)*B*x,     A*Bx=(lambda)*x,     or
                 B*A*x=(lambda)*x

       chegvd  - compute all the eigenvalues, and optionally, the eigenvectors
                 of a complex generalized Hermitian-definite eigenproblem,  of
                 the     form     A*x=(lambda)*B*x,     A*Bx=(lambda)*x,    or
                 B*A*x=(lambda)*x

       chegvx - compute selected eigenvalues, and optionally, eigenvectors  of
                 a complex generalized Hermitian-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       chemm - perform one of the matrix-matrix operations C  :=  alpha*A*B  +
                 beta*C or C := alpha*B*A + beta*C

       chemv - perform the matrix-vector operationy := alpha*A*x + beta*y

       cher  - perform the hermitian rank 1 operation   A := alpha*x*conjg( x'
                 ) + A

       cher2 - perform the hermitian rank 2 operation   A := alpha*x*conjg( y'
                 ) + conjg( alpha )*y*conjg( x' ) + A

       cher2k  -  perform  one  of  the  Hermitian  rank  2k operations   C :=
                 alpha*A*conjg( B' ) + conjg( alpha )*B*conjg( A' )  +  beta*C
                 or  C := alpha*conjg( A' )*B + conjg( alpha )*conjg( B' )*A +
                 beta*C

       cherfs - improve the computed solution to a system of linear  equations
                 when the coefficient matrix is Hermitian indefinite, and pro-
                 vide error bounds and backward error estimates for the  solu-
                 tion

       cherfsx - improve the computed solution to a system of linear equations
                 when the coefficient matrix is Hermitian indefinite, and pro-
                 vide  error bounds and backward error estimates for the solu-
                 tion

       cherk -  perform  one  of  the  Hermitian  rank  k  operations    C  :=
                 alpha*A*conjg(  A'  )  + beta*C or C := alpha*conjg( A' )*A +
                 beta*C

       chesv - compute the solution to a complex system  of  linear  equations
                 A*X=B,  where A is an N-by-N Hermitian matrix and X and B are
                 N-by-NRHS matrices

       chesv_rook - compute the solution to a system of linear equations A*X=B
                 for   Hermitian  matrices  using  the  bounded  Bunch-Kaufman
                 ("rook") diagonal pivoting method

       chesvx - use the diagonal pivoting factorization to compute  the  solu-
                 tion to a complex system of linear equations A*X = B, where A
                 is an N-by-N Hermitian matrix  and  X  and  B  are  N-by-NRHS
                 matrices

       chesvxx  -  compute  the solution to system of linear equations A*X = B
                 for Hermitian matrices

       chetd2 - reduce a Hermitian matrix to real symmetric  tridiagonal  form
                 by an unitary similarity transformation (unblocked algorithm)

       chetf2 - compute the factorization of a complex Hermitian matrix, using
                 the  diagonal  pivoting  method  (unblocked algorithm calling
                 Level 2 BLAS)

       chetf2_rook - compute the factorization of a complex Hermitian  indefi-
                 nite matrix using the bounded Bunch-Kaufman ("rook") diagonal
                 pivoting method (unblocked algorithm)

       chetrd - reduce a complex Hermitian matrix A to real symmetric tridiag-
                 onal form T by a unitary similarity transformation

       chetrf  -  compute  the  factorization  of a complex Hermitian matrix A
                 using the Bunch-Kaufman diagonal pivoting method

       chetrf_rook - compute the factorization of a complex Hermitian  indefi-
                 nite matrix using the bounded Bunch-Kaufman ("rook") diagonal
                 pivoting method (blocked algorithm, calling Level 3 BLAS)

       chetri - compute the inverse of a complex Hermitian indefinite matrix A
                 using the factorization A = U*D*U**H or A = L*D*L**H computed
                 by CHETRF

       chetri2 - compute the inverse of a COMPLEX Hermitian indefinite  matrix
                 A  using  the factorization A=U*D*U**T or A=L*D*L**T computed
                 by CHETRF

       chetri2x - computes the  inverse  of  a  complex  Hermitian  indefinite
                 matrix A using the factorization A = U*D*U**H or A = L*D*L**H
                 computed by CHETRF

       chetri_rook - compute the inverse of a Hermitian matrix using the  fac-
                 torization  obtained  with the bounded Bunch-Kaufman ("rook")
                 diagonal pivoting method

       chetrs - solve a system of linear equations A*X = B with a complex Her-
                 mitian  matrix  A using the factorization A = U*D*U**H or A =
                 L*D*L**H computed by CHETRF

       chetrs2 - solve a system of linear equations A*X =  B  with  a  complex
                 Hermitian  matrix A using the factorization A = U*D*U**H or A
                 = L*D*L**H computed by CHETRF and converted by CSYCONV

       chetrs_rook - compute the solution to  a  system  of  linear  equations
                 A*X=B  for  Hermitian  matrices  using factorization obtained
                 with one of the bounded  diagonal  pivoting  methods  (max  2
                 interchanges)

       chfrk - perform a Hermitian rank-k operation for matrix in RFP format

       chgeqz  - implement a single-shift version of the QZ method for finding
                 the  generalized  eigenvalues  w(i)=ALPHA(i)/BETA(i)  of  the
                 equation    det(  A-w(i)  B  ) = 0  If JOB='S', then the pair
                 (A,B) is simultaneously reduced to Schur form (i.e., A and  B
                 are both upper triangular) by applying one unitary tranforma-
                 tion (usually called Q) on  the  left  and  another  (usually
                 called Z) on the right

       chla_transtype  -  translate from a BLAST-specified integer constant to
                 the character string specifying a transposition operation

       chpcon - estimate the reciprocal of the condition number of  a  complex
                 Hermitian  packed  matrix  A  using  the  factorization  A  =
                 U*D*U**H or A = L*D*L**H computed by CHPTRF

       chpev - compute all the eigenvalues and, optionally, eigenvectors of  a
                 complex Hermitian matrix in packed storage

       chpevd - compute all the eigenvalues and, optionally, eigenvectors of a
                 complex Hermitian matrix A in packed storage

       chpevx - compute selected eigenvalues and, optionally, eigenvectors  of
                 a complex Hermitian matrix A in packed storage

       chpgst  -  reduce a complex Hermitian-definite generalized eigenproblem
                 to standard form, using packed storage

       chpgv - compute all the eigenvalues and, optionally,  the  eigenvectors
                 of  a complex generalized Hermitian-definite eigenproblem, of
                 the    form     A*x=(lambda)*B*x,     A*Bx=(lambda)*x,     or
                 B*A*x=(lambda)*x

       chpgvd  - compute all the eigenvalues and, optionally, the eigenvectors
                 of a complex generalized Hermitian-definite eigenproblem,  of
                 the     form     A*x=(lambda)*B*x,     A*Bx=(lambda)*x,    or
                 B*A*x=(lambda)*x

       chpgvx - compute selected eigenvalues and, optionally, eigenvectors  of
                 a complex generalized Hermitian-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       chpmv - perform the matrix-vector operationy := alpha*A*x + beta*y

       chpr - perform the hermitian rank 1 operation   A := alpha*x*conjg(  x'
                 ) + A

       chpr2 - perform the Hermitian rank 2 operation   A := alpha*x*conjg( y'
                 ) + conjg( alpha )*y*conjg( x' ) + A

       chprfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix  is  Hermitian  indefinite and
                 packed, and provide error bounds and backward error estimates
                 for the solution

       chpsv - compute the solution to a complex system of linear equationsA*X
                 = B, where A is an N-by-N Hermitian matrix stored  in  packed
                 format and X and B are N-by-NRHS matrices

       chpsvx  -  use  the diagonal pivoting factorization A = U*D*U**H or A =
                 L*D*L**H to compute the solution to a complex system of  lin-
                 ear  equations  A  *  X  =  B, where A is an N-by-N Hermitian
                 matrix stored in packed format and  X  and  B  are  N-by-NRHS
                 matrices

       chptrd  -  reduce a complex Hermitian matrix A stored in packed form to
                 real symmetric tridiagonal form T  by  a  unitary  similarity
                 transformation

       chptrf - compute the factorization of a complex Hermitian packed matrix
                 A using the Bunch-Kaufman diagonal pivoting method

       chptri - compute the inverse of a complex Hermitian indefinite matrix A
                 in packed storage using the factorization A = U*D*U**H or A =
                 L*D*L**H computed by CHPTRF

       chptrs - solve a system of linear equations A*X = B with a complex Her-
                 mitian  matrix A stored in packed format using the factoriza-
                 tion A = U*D*U**H or A = L*D*L**H computed by CHPTRF

       chsein - use inverse iteration to  find  specified  right  and/or  left
                 eigenvectors of a complex upper Hessenberg matrix H

       chseqr  -  compute the eigenvalues of a complex upper Hessenberg matrix
                 H, and, optionally, the matrices  T  and  Z  from  the  Schur
                 decomposition  H  =  Z T Z**H, where T is an upper triangular
                 matrix (the Schur form), and Z is the unitary matrix of Schur
                 vectors

       cinfinite_norm_error  -  A  utility function of the SuperLU solver that
                 computes the infinity-norm of an array of  vectors  that  are
                 approximations to the exact solution vector.

       cjadmm - Jagged diagonal matrix-matrix multiply (modified Ellpack)

       cjadrp - right permutation of a jagged diagonal matrix

       cjadsm - Jagged-diagonal format triangular solve

       cla_gbamv - perform a matrix-vector operation to calculate error bounds

       cla_gbrcond_c  -  compute  the  infinity  norm  condition   number   of
                 op(A)*inv(diag(c)) for general banded matrices

       cla_gbrcond_x   -   compute  the  infinity  norm  condition  number  of
                 op(A)*diag(x) for general banded matrices

       cla_gbrfsx_extended - improve the computed solution to a system of lin-
                 ear  equations  for  general  banded  matrices  by performing
                 extra-precise iterative refinement and provide  error  bounds
                 and backward error estimates for the solution

       cla_gbrpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U) for a general banded matrix

       cla_geamv - compute a matrix-vector product using a general  matrix  to
                 calculate error bounds

       cla_gercond_c   -   compute  the  infinity  norm  condition  number  of
                 op(A)*inv(diag(c)) for general matrices

       cla_gercond_x  -  compute  the  infinity  norm  condition   number   of
                 op(A)*diag(x) for general matrices

       cla_gerfsx_extended - improve the computed solution to a system of lin-
                 ear equations by performing extra-precise  iterative  refine-
                 ment  and  provide  error bounds and backward error estimates
                 for the solution

       cla_gerpvgrw - compute the reciprocal pivot  growth  factor  using  the
                 "max absolute element" norm

       cla_heamv  -  compute a matrix-vector product using a Hermitian indefi-
                 nite matrix to calculate error bounds

       cla_hercond_c  -  compute  the  infinity  norm  condition   number   of
                 op(A)*inv(diag(c)) for Hermitian indefinite matrices

       cla_hercond_x   -   compute  the  infinity  norm  condition  number  of
                 op(A)*diag(x) for Hermitian indefinite matrices

       cla_herfsx_extended - improve the computed solution to a system of lin-
                 ear equations for Hermitian indefinite matrices by performing
                 extra-precise iterative refinement and provide  error  bounds
                 and backward error estimates for the solution

       cla_herpvgrw  -  compute  the  reciprocal pivot growth factor using the
                 "max absolute element" norm

       cla_lin_berr - compute a component-wise relative backward error

       cla_porcond_c  -  compute  the  infinity  norm  condition   number   of
                 op(A)*inv(diag(c)) for Hermitian positive-definite matrices

       cla_porcond_x   -   compute  the  infinity  norm  condition  number  of
                 op(A)*diag(x) for Hermitian positive-definite matrices

       cla_porfsx_extended - improve the computed solution to a system of lin-
                 ear  equations  for  symmetric or Hermitian positive-definite
                 matrices by performing extra-precise iterative refinement and
                 provide  error  bounds  and  backward error estimates for the
                 solution

       cla_porpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U)  for  a symmetric or Hermitian positive-defi-
                 nite matrix

       cla_syamv - compute a matrix-vector product using a  symmetric  indefi-
                 nite matrix to calculate error bounds

       cla_syrcond_c   -   compute  the  infinity  norm  condition  number  of
                 op(A)*inv(diag(c)) for symmetric indefinite matrices

       cla_syrcond_x  -  compute  the  infinity  norm  condition   number   of
                 op(A)*diag(x) for symmetric indefinite matrices

       cla_syrfsx_extended - improve the computed solution to a system of lin-
                 ear equations for symmetric indefinite matrices by performing
                 extra-precise  iterative  refinement and provide error bounds
                 and backward error estimates for the solution

       cla_syrpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U) for a symmetric indefinite matrix

       cla_wwaddw - add a vector into a doubled-single vector

       claed0  -  compute all eigenvalues and corresponding eigenvectors of an
                 unreduced symmetric tridiagonal matrix using the  divide  and
                 conquer method. Used by cstedc

       claed7  -  compute  the  updated eigensystem of a diagonal matrix after
                 modification by a rank-one symmetric matrix. Used by  cstedc,
                 when the original matrix is dense

       claed8  - merge eigenvalues and deflates secular equation. Used by cst-
                 edc, when the original matrix is dense

       clahef - compute a partial factorization of a complex Hermitian indefi-
                 nite  matrix using the Bunch-Kaufman diagonal pivoting method
                 (blocked algorithm, calling Level 3 BLAS)

       clahef_rook - compute a partial factorization of  a  complex  Hermitian
                 indefinite  matrix  using  the bounded Bunch-Kaufman ("rook")
                 diagonal pivoting method (blocked algorithm, calling Level  3
                 BLAS)

       clals0  -  apply  back multiplying factors in solving the least squares
                 problem using divide and conquer SVD approach. Used by cgelsd

       clalsa  -  compute  the  SVD of the coefficient matrix in compact form.
                 Used by cgelsd

       clalsd - use the singular value decomposition of A to solve  the  least
                 squares problem

       clangs  -  returns the value of the one-norm, or the Frobenius-norm, or
                 the infinity-norm, or the element with largest absolute value
                 of a general real matrix A in sparse format.

       clanhf  - return the value of the 1-norm, or the Frobenius norm, or the
                 infinity norm, or the element of largest absolute value of  a
                 Hermitian matrix in RFP format

       claqgs  -  a SuperLU function that equilibrates a general sparse M by N
                 matrix A.

       clarscl2 - perform reciprocal diagonal scaling on a vector

       clarz - apply a complex elementary reflector  H  to  a  complex  M-by-N
                 matrix C, from either the left or the right

       clarzb  -  apply a complex block reflector H or its transpose H**H to a
                 complex distributed M-by-N C from the left or the right

       clarzt - form the triangular factor T of a complex block reflector H of
                 order  >  n,  which  is  defined as a product of k elementary
                 reflectors

       clascl2 - perform diagonal scaling on a vector

       clasyf - compute a partial factorization of a complex symmetric  matrix
                 using the Bunch-Kaufman diagonal pivoting method

       clasyf_rook  -  compute  a partial factorization of a complex symmetric
                 matrix using the bounded Bunch-Kaufman ("rook") diagonal piv-
                 oting method

       clatrz  - factor an upper trapezoidal matrix by means of unitary trans-
                 formations

       clatzm - routine is deprecated and has been replaced by routine CUNMRZ

       cosqb - synthesize a Fourier sequence from its representation in  terms
                 of a cosine series with odd wave numbers. The COSQ operations
                 are unnormalized inverses of themselves, so a call  to  COSQF
                 followed  by a call to COSQB will multiply the input sequence
                 by 4 * N.

       cosqf - compute the Fourier coefficients in a cosine series representa-
                 tion  with  only  odd  wave  numbers. The COSQ operations are
                 unnormalized inverses of themselves, so a call to COSQF  fol-
                 lowed  by a call to COSQB will multiply the input sequence by
                 4 * N.

       cosqi - initialize the array WSAVE, which is used  in  both  COSQF  and
                 COSQB.

       cost  -  compute  the  discrete  Fourier  cosine  transform  of an even
                 sequence.  The COST transforms are unnormalized  inverses  of
                 themselves,  so  a  call  of COST followed by another call of
                 COST will multiply the input sequence by 2 * (N-1).

       costi - initialize the array WSAVE, which is used in COST.

       cpbcon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm)  of a complex Hermitian positive definite band matrix
                 using the Cholesky factorization A = U**H*U  or  A  =  L*L**H
                 computed by CPBTRF

       cpbequ - compute row and column scalings intended to equilibrate a Her-
                 mitian positive definite band matrix A and reduce its  condi-
                 tion number (with respect to the two-norm)

       cpbrfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is  Hermitian  positive  definite
                 and banded, provide error bounds and backward error estimates
                 for the solution

       cpbstf - compute a split Cholesky factorization of a complex  Hermitian
                 positive definite band matrix A

       cpbsv  -  compute  the solution to a complex system of linear equations
                 A*X=B, where A is an N-by-N Hermitian positive definite  band
                 matrix and X and B are N-by-NRHS matrices

       cpbsvx  -  use  the Cholesky factorization to compute the solution to a
                 complex system of linear equations A*X=B, where A is an N-by-
                 N  Hermitian positive definite band matrix and X and B are N-
                 by-NRHS matrices

       cpbtf2 - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite band matrix A

       cpbtrf - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite band matrix A

       cpbtrs - solve a system of linear equations A*X = B  with  a  Hermitian
                 positive definite band matrix A using the Cholesky factoriza-
                 tion A = U**H*U or A = L*L**H computed by CPBTRF

       cpftrf - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite matrix A, the block version of the algorithm

       cpftri  -  compute the inverse of a complex Hermitian positive definite
                 matrix A using the Cholesky factorization computed by CPFTRF

       cpftrs - solve a system of linear equations A*X = B  with  a  Hermitian
                 positive  definite  matrix A using the Cholesky factorization
                 computed by CPFTRF

       cpocon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm) of a complex Hermitian positive definite matrix using
                 the Cholesky factorization A = U**H*U or A = L*L**H  computed
                 by CPOTRF

       cpoequ - compute row and column scalings intended to equilibrate a Her-
                 mitian positive definite matrix A and  reduce  its  condition
                 number (with respect to the two-norm)

       cpoequb  -  compute  row  and column scalings intended to equilibrate a
                 symmetric positive definite matrix A and reduce its condition
                 number with respect to the two-norm

       cporfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is Hermitian  positive  definite,
                 provide  error  bounds  and  backward error estimates for the
                 solution

       cporfsx - improve the computed solution to a system of linear equations
                 when  the  coefficient matrix is symmetric positive definite,
                 provide error bounds and backward  error  estimates  for  the
                 solution

       cposv - compute the solution to a complex system of linear equationsA*X
                 = B, where A is an N-by-N Hermitian positive definite  matrix
                 and X and B are N-by-NRHS matrices

       cposvx  -  use  the Cholesky factorization to compute the solution to a
                 complex system of linear equations  A*X = B, where A is an N-
                 by-N Hermitian positive definite matrix and X and B are N-by-
                 NRHS matrices

       cposvxx - compute the solution to a complex system of linear  equations
                 A*X  =  B,  where  A is an N-by-N symmetric positive definite
                 matrix and X and B are N-by-NRHS matrices

       cpotf2 - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite matrix A

       cpotrf - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite matrix A

       cpotri - compute the inverse of a complex Hermitian  positive  definite
                 matrix  A  using the Cholesky factorization A = U**H*U or A =
                 L*L**H computed by CPOTRF

       cpotrs - solve a system of linear equations A*X = B  with  a  Hermitian
                 positive definite matrix A using the Cholesky factorization A
                 = U**H*U or A = L*L**H computed by CPOTRF

       cppcon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm)  of  a  complex  Hermitian  positive  definite packed
                 matrix using the Cholesky factorization A =  U**H*U  or  A  =
                 L*L**H computed by CPPTRF

       cppequ - compute row and column scalings intended to equilibrate a Her-
                 mitian positive definite  matrix  A  in  packed  storage  and
                 reduce its condition number (with respect to the two-norm)

       cpprfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is  Hermitian  positive  definite
                 and packed, provide error bounds and backward error estimates
                 for the solution

       cppsv - compute the solution to a complex system  of  linear  equations
                 A*X  =  B,  where  A is an N-by-N Hermitian positive definite
                 matrix stored in packed format and  X  and  B  are  N-by-NRHS
                 matrices

       cppsvx  -  use  the Cholesky factorization to compute the solution to a
                 complex system of linear equations  A*X = B, where A is an N-
                 by-N Hermitian positive definite matrix stored in packed for-
                 mat and X and B are N-by-NRHS matrices

       cpptrf - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite matrix A stored in packed format

       cpptri  -  compute the inverse of a complex Hermitian positive definite
                 matrix A using the Cholesky factorization A = U**H*U or  A  =
                 L*L**H computed by CPPTRF

       cpptrs  -  solve  a system of linear equations A*X = B with a Hermitian
                 positive definite  matrix  A  in  packed  storage  using  the
                 Cholesky  factorization  A = U**H*U or A = L*L**H computed by
                 CPPTRF

       cpstf2 - compute the Cholesky factorization with complete pivoting of a
                 complex Hermitian positive semidefinite matrix A

       cpstrf - compute the Cholesky factorization with complete pivoting of a
                 complex Hermitian positive semidefinite matrix A

       cptcon - compute the reciprocal of the condition number (in the 1-norm)
                 of  a  complex Hermitian positive definite tridiagonal matrix
                 using the factorization A = L*D*L**H or A = U**H*D*U computed
                 by CPTTRF

       cpteqr  -  compute  all  eigenvalues and, optionally, eigenvectors of a
                 symmetric positive definite tridiagonal matrix by first  fac-
                 toring  the  matrix  using  SPTTRF and then calling CBDSQR to
                 compute the singular values of the bidiagonal factor

       cptrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix is Hermitian positive definite
                 and tridiagonal, provide  error  bounds  and  backward  error
                 estimates for the solution

       cptsv  -  compute  the solution to a complex system of linear equations
                 A*X = B, where A is an  N-by-N  Hermitian  positive  definite
                 tridiagonal matrix, and X and B are N-by-NRHS matrices

       cptsvx  - use the factorization A = L*D*L**H to compute the solution to
                 a complex system of linear equations A*X = B, where A  is  an
                 N-by-N  Hermitian  positive definite tridiagonal matrix and X
                 and B are N-by-NRHS matrices

       cpttrf - compute the L*D*L' factorization of a complex Hermitian  posi-
                 tive definite tridiagonal matrix A

       cpttrs  -  solve  a tridiagonal system of the form  A * X = B using the
                 factorization A = U'*D*U or A = L*D*L' computed by CPTTRF

       cptts2 - solve a tridiagonal system of the form  A * X =  B  using  the
                 factorization A = U'*D*U or A = L*D*L' computed by CPTTRF

       crot  -  apply  a plane rotation, where the cos (C) is real and the sin
                 (S) is complex, and the vectors X and Y are complex

       crotg - Construct a Given's plane rotation

       cscal - Compute y := alpha * y

       csctr - Scatters elements from x into y

       cskymm - Skyline format matrix-matrix multiply

       cskysm - Skyline format triangular solve

       cspcon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm) of a complex symmetric packed matrix A using the fac-
                 torization A = U*D*U**T or A = L*D*L**T computed by CSPTRF

       csprfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix  is  symmetric  indefinite and
                 packed, provide error bounds and backward error estimates for
                 the solution

       cspsv - compute the solution to a complex system of linear equationsA*X
                 = B, where A is an N-by-N symmetric matrix stored  in  packed
                 format and X and B are N-by-NRHS matrices

       cspsvx  -  use  the diagonal pivoting factorization A = U*D*U**T or A =
                 L*D*L**T to compute the solution to a complex system of  lin-
                 ear  equations  A  *  X  =  B, where A is an N-by-N symmetric
                 matrix stored in packed format and  X  and  B  are  N-by-NRHS
                 matrices

       csptrf  -  compute  the  factorization  of a complex symmetric matrix A
                 stored in packed format using the Bunch-Kaufman diagonal piv-
                 oting method

       csptri - compute the inverse of a complex symmetric indefinite matrix A
                 in packed storage using the factorization A = U*D*U**T or A =
                 L*D*L**T computed by CSPTRF

       csptrs - solve a system of linear equations A*X = B with a complex sym-
                 metric matrix A stored in packed format using the  factoriza-
                 tion A = U*D*U**T or A = L*D*L**T computed by CSPTRF

       csrot - Apply a plane rotation

       csscal - Compute y := alpha * y

       cstedc  -  compute  all  eigenvalues and, optionally, eigenvectors of a
                 symmetric tridiagonal matrix using  the  divide  and  conquer
                 method

       cstegr - Compute T-sigma_i = L_i D_i L_i^T, such that L_i D_i L_i^T  is
                 a relatively robust representation

       cstein - compute the  eigenvectors  of  a  real  symmetric  tridiagonal
                 matrix   T  corresponding  to  specified  eigenvalues,  using
                 inverse iteration

       cstemr - compute selected eigenvalues and, optionally, eigenvectors  of
                 a real symmetric tridiagonal matrix T

       csteqr  -  compute  all  eigenvalues and, optionally, eigenvectors of a
                 symmetric tridiagonal matrix using  the  implicit  QL  or  QR
                 method

       cstsv  - compute the solution to a complex system of linear equations A
                 * X = B where A is a symmetric tridiagonal matrix

       csttrf - compute the factorization of a complex  symmetric  tridiagonal
                 matrix A using the Bunch-Kaufman diagonal pivoting method

       csttrs  -  compute the solution to a complex system of linear equations
                 A*X = B, where A is an N-by-N  symmetric  tridiagonal  matrix
                 and X and B are N-by-NRHS matrices

       cswap - Exchange vectors x and y

       csycon  -  estimate  the  reciprocal  of  the  condition number (in the
                 1-norm) of a complex symmetric matrix A using the  factoriza-
                 tion A = U*D*U**T or A = L*D*L**T computed by CSYTRF

       csycon_rook  -  estimate the reciprocal of the condition number (in the
                 1-norm) of a complex symmetric matrix A using the  factoriza-
                 tion A = U*D*U**T or A = L*D*L**T computed by CSYTRF_ROOK

       csyconv - convert A given by TRF into L and D and vice-versa

       csyequb  -  compute  row  and column scalings intended to equilibrate a
                 symmetric matrix A  and  reduce  its  condition  number  with
                 respect to the two-norm

       csymm  -  perform  one  of the matrix-matrix operationsC := alpha*A*B +
                 beta*C or C := alpha*B*A + beta*C

       csyr2k - perform one  of  the  symmetric  rank  2k  operations    C  :=
                 alpha*A*B'  +  alpha*B*A'  +  beta*C  or  C  :=  alpha*A'*B +
                 alpha*B'*A + beta*C

       csyrfs - improve the computed solution to a system of linear  equations
                 when  the coefficient matrix is symmetric indefinite, provide
                 error bounds and backward error estimates for the solution

       csyrfsx - improve the computed solution to a system of linear equations
                 when  the coefficient matrix is symmetric indefinite, provide
                 error bounds and backward error estimates for the solution

       csyrk -  perform  one  of  the  symmetric  rank  k  operations    C  :=
                 alpha*A*A' + beta*C or C := alpha*A'*A + beta*C

       csysv - compute the solution to a complex system of linear equationsA*X
                 = B, where A is an N-by-N symmetric matrix and X and B are N-
                 by-NRHS matrices

       csysv_rook - compute the solution to system of linear equations A*X = B
                 for symmetric matrices. CSYTRF_ROOK is called to compute  the
                 factorization of A

       csysvx  -  use the diagonal pivoting factorization to compute the solu-
                 tion to a complex system of linear equations A*X = B, where A
                 is  an  N-by-N  symmetric  matrix  and  X and B are N-by-NRHS
                 matrices

       csysvxx - compute the solution to complex system  of  linear  equations
                 A*X = B for symmetric matrices

       csytf2  -  compute  the  factorization  of a complex symmetric matrix A
                 using the Bunch-Kaufman diagonal pivoting method

       csytf2_rook - compute the factorization of a complex symmetric  indefi-
                 nite matrix using the bounded Bunch-Kaufman ("rook") diagonal
                 pivoting method (unblocked algorithm)

       csytrf - compute the factorization of  a  complex  symmetric  matrix  A
                 using the Bunch-Kaufman diagonal pivoting method

       csytrf_rook  -  compute the factorization of a complex symmetric matrix
                 using the bounded Bunch-Kaufman  ("rook")  diagonal  pivoting
                 method (blocked algorithm)

       csytri - compute the inverse of a complex symmetric indefinite matrix A
                 using the factorization A = U*D*U**T or A = L*D*L**T computed
                 by CSYTRF

       csytri2  - compute the inverse of a COMPLEX symmetric indefinite matrix
                 A using the factorization A = U*D*U**T or A =  L*D*L**T  com-
                 puted by CSYTRF

       csytri2x  - compute the inverse of a real symmetric indefinite matrix A
                 using the factorization computed by CSYTRF

       csytrii_rook - compute the inverse of a  complex  symmetric  indefinite
                 matrix A using the factorization A = U*D*U**T or A = L*D*L**T
                 computed by CSYTRF_ROOK

       csytrs - solve a system of linear equations A*X = B with a complex sym-
                 metric  matrix  A using the factorization A = U*D*U**T or A =
                 L*D*L**T computed by CSYTRF

       csytrs2 - solve a system of linear equations A*X =  B  with  a  complex
                 symmetric matrix A using the factorization computed by CSYTRF
                 and converted by CSYCONV

       csytrs_rook - solve a system of linear equations A*X = B with a complex
                 symmetric  matrix A using the factorization A = U*D*U**T or A
                 = L*D*L**T computed by CSYTRF_ROOK

       ctbcon - estimate the reciprocal of the condition number of a  triangu-
                 lar band matrix A, in either the 1-norm or the infinity-norm

       ctbmv  -  perform  one of the matrix-vector operationsx := A*x, or x :=
                 A'*x, or x := conjg( A' )*x

       ctbrfs - provide error bounds and  backward  error  estimates  for  the
                 solution  to  a  system of linear equations with a triangular
                 band coefficient matrix

       ctbsv - solve one of the systems of equations   A*x = b, or A'*x  =  b,
                 or conjg( A' )*x = b

       ctbtrs  - solve a triangular system of the form A*X = B, A**T*X = B, or
                 A**H*X = B, where A is a triangular band matrix of  order  N,
                 and B is an N-by-NRHS matrix

       ctfsm  - solve a matrix equation (one operand is a triangular matrix in
                 RFP format)

       ctftri - compute the inverse of a triangular matrix  A  stored  in  RFP
                 format

       ctfttp - copy a triangular matrix from the rectangular full packed for-
                 mat (TF) to the standard packed format (TP)

       ctfttr - copy a triangular matrix from the rectangular full packed for-
                 mat (TF) to the standard full format (TR)

       ctgevc  -  compute  the right and/or left generalized eigenvectors of a
                 pair of complex upper triangular matrices with real  diagonal
                 elements (A,B) obtained from the generalized Schur factoriza-
                 tion of an original pair  of  complex  nonsymmetric  matrices
                 (AO,BO)

       ctgexc  -  reorder  the  generalized  Schur  decomposition of a complex
                 matrix pair using an orthogonal or unitary equivalence trans-
                 formation

       ctgsen  -  reorder  the  generalized  Schur  decomposition of a complex
                 matrix pair (A, B), so that a selected cluster of eigenvalues
                 appears in the leading diagonal blocks of the pair (A,B)

       ctgsja - compute the generalized singular value decomposition (GSVD) of
                 two complex upper triangular (or trapezoidal) matrices A  and
                 B

       ctgsna  - estimate reciprocal condition numbers for specified eigenval-
                 ues and/or eigenvectors of a matrix pair (A, B)

       ctgsyl - solve the generalized Sylvester equation

       ctpcon - estimate the reciprocal of the condition number  of  a  packed
                 triangular  matrix  A,  in either the 1-norm or the infinity-
                 norm

       ctpmqrt - apply a complex orthogonal matrix Q obtained from a "triangu-
                 lar-pentagonal"  complex  block reflector H to a general com-
                 plex matrix C, which consists of two blocks

       ctpmv - perform one of the matrix-vector operationsx := A*x,  or  x  :=
                 A'*x, or x := conjg( A' )*x

       ctpqrt  -  compute a blocked QR factorization of a complex "triangular-
                 pentagonal" matrix C, which is composed of a triangular block
                 A and pentagonal block B, using the compact WY representation
                 for Q

       ctpqrt2 - compute a QR factorization of a real or complex  "triangular-
                 pentagonal"  matrix,  which is composed of a triangular block
                 and a pentagonal block, using the compact  WY  representation
                 for Q

       ctprfs  -  provide  error  bounds  and backward error estimates for the
                 solution to a system of linear equations  with  a  triangular
                 packed coefficient matrix

       ctpsv  -  solve one of the systems of equations   A*x = b, or A'*x = b,
                 or conjg( A' )*x = b

       ctptri - compute the inverse of a complex  upper  or  lower  triangular
                 matrix A stored in packed format

       ctptrs  - solve a triangular system of the form A*X = B, A**T*X = B, or
                 A**H*X = B, where A is a triangular matrix of order N  stored
                 in packed format, and B is an N-by-NRHS matrix

       ctpttf  - copy a triangular matrix from the standard packed format (TP)
                 to the rectangular full packed format (TF)

       ctpttr - copy a triangular matrix from the standard packed format  (TP)
                 to the standard full format (TR)

       ctrans - transpose and scale source matrix

       ctrcon  - estimate the reciprocal of the condition number of a triangu-
                 lar matrix A, in either the 1-norm or the infinity-norm

       ctrevc - compute some or all of the right and/or left eigenvectors of a
                 complex upper triangular matrix T

       ctrexc  -  reorder  the  Schur  factorization  of  a complex matrix A =
                 Q*T*Q**H, so that the diagonal element of T  with  row  index
                 IFST is moved to row ILST

       ctrmm  -  perform  one  of the matrix-matrix operationsB := alpha*op( A
                 )*B, or B := alpha*B*op( A )  where alpha is a scalar,  B  is
                 an  m  by  n matrix, A is a unit, or non-unit, upper or lower
                 triangular matrix and op( A ) is one of   op( A ) = A or  op(
                 A ) = A' or op( A ) = conjg( A' )

       ctrmv  -  perform  one of the matrix-vector operationsx := A*x, or x :=
                 A'*x, or x := conjg( A' )*x

       ctrrfs - provide error bounds and  backward  error  estimates  for  the
                 solution  to  a  system of linear equations with a triangular
                 coefficient matrix

       ctrsen - reorder the Schur  factorization  of  a  complex  matrix  A  =
                 Q*T*Q**H,  so  that a selected cluster of eigenvalues appears
                 in the leading positions on the diagonal of the upper  trian-
                 gular  matrix  T,  and  the  leading  columns  of  Q  form an
                 orthonormal basis of the corresponding right  invariant  sub-
                 space

       ctrsm - solve one of the matrix equations op( A )*X = alpha*B, or X*op(
                 A ) = alpha*B

       ctrsna - estimate reciprocal condition numbers for specified  eigenval-
                 ues  and/or  right eigenvectors of a complex upper triangular
                 matrix T (or of any matrix Q*T*Q**H with Q unitary)

       ctrsv - solve one of the systems of equations   A*x = b, or A'*x  =  b,
                 or conjg( A' )*x = b

       ctrsyl - solve the complex Sylvester matrix equation

       ctrti2  -  compute  the  inverse of a complex upper or lower triangular
                 matrix

       ctrtri - compute the inverse of a complex  upper  or  lower  triangular
                 matrix A

       ctrtrs  - solve a triangular system of the form A*X = B, A**T*X = B, or
                 A**H*X = B, where A is a triangular matrix of order N, and  B
                 is an N-by-NRHS matrix

       ctrttf - copy a triangular matrix from the standard full format (TR) to
                 the rectangular full packed format (TF)

       ctrttp - copy a triangular matrix from the standard full format (TR) to
                 the standard packed format (TP)

       ctzrqf - routine is deprecated and has been replaced by routine CTZRZF

       ctzrzf  - reduce the M-by-N ( M<=N ) complex upper trapezoidal matrix A
                 to upper triangular form by means of unitary transformations

       cunbdb - simultaneously bidiagonalizes the blocks of an  M-by-M  parti-
                 tioned unitary matrix

       cunbdb1  - simultaneously bidiagonalize the blocks of a tall and skinny
                 matrix with orthonomal columns

       cunbdb2 - simultaneously bidiagonalize the blocks of a tall and  skinny
                 matrix with orthonomal columns

       cunbdb3  - simultaneously bidiagonalize the blocks of a tall and skinny
                 matrix with orthonomal columns

       cunbdb4 - simultaneously bidiagonalize the blocks of a tall and  skinny
                 matrix with orthonomal columns

       cunbdb5 - orthogonalize one column vector with respect to the orthonor-
                 mal columns of another column vector

       cunbdb6 - orthogonalize one column vector with respect to the orthonor-
                 mal columns of another column vector

       cuncsd  - compute the CS decomposition of an M-by-M partitioned unitary
                 matrix

       cuncsd2by1 - compute the CS decomposition  of  an  M-by-Q  matrix  with
                 orthonormal  columns  that has been partitioned into a 2-by-1
                 block structure

       cung2l - generate all or part of the unitary matrix Q from a QL factor-
                 ization determined by cgeqlf (unblocked algorithm)

       cung2r  - generate an M-by-N complex matrix Q with orthonormal columns,
                 which is defined as the first N columns of  a  product  of  K
                 elementary reflectors of order M

       cungbr  - generate one of the complex unitary matrices Q or P**H deter-
                 mined by CGEBRD when reducing a complex matrix A to  bidiago-
                 nal form

       cunghr  -  generate  a complex unitary matrix Q which is defined as the
                 product of IHI-ILO  elementary  reflectors  of  order  N,  as
                 returned by CGEHRD

       cungl2  -  generate all or part of the unitary matrix Q from an LQ fac-
                 torization determined by cgelqf (unblocked algorithm)

       cunglq - generate an M-by-N complex matrix  Q  with  orthonormal  rows,
                 which  is  defined as the first M rows of a product of K ele-
                 mentary reflectors of order N

       cungql - generate an M-by-N complex matrix Q with orthonormal  columns,
                 which is defined as the last N columns of a product of K ele-
                 mentary reflectors of order M

       cungqr - generate an M-by-N complex matrix Q with orthonormal  columns,
                 which  is  defined  as  the first N columns of a product of K
                 elementary reflectors of order M

       cungr2 - generate all or part of the unitary matrix Q from an  RQ  fac-
                 torization determined by cgerqf (unblocked algorithm)

       cungrq  -  generate  an  M-by-N complex matrix Q with orthonormal rows,
                 which is defined as the last M rows of a product of K elemen-
                 tary reflectors of order N

       cungtr  -  generate  a complex unitary matrix Q which is defined as the
                 product of n-1 elementary reflectors of order N, as  returned
                 by CHETRD

       cunm2l - multiply a general matrix by the unitary matrix from a QL fac-
                 torization determined by cgeqlf (unblocked algorithm)

       cunm2r - multiply a general matrix by the unitary matrix from a QR fac-
                 torization determined by cgeqrf (unblocked algorithm)

       cunmbr  -  overwrite  the  general  complex M-by-N matrix C with Q*C or
                 Q**H*C or C*Q**H or C*Q or P*C or P**H*C or C*P or C*P**H

       cunmhr - overwrite the general complex M-by-N  matrix  C  with  Q*C  or
                 Q**H*C or C*Q**H or C*Q

       cunml2 - multiply a general matrix by the unitary matrix from a LQ fac-
                 torization determined by cgelqf (unblocked algorithm)

       cunmlq - overwrite the general complex M-by-N matrix  C  with  Q*C,  or
                 Q**H*C,  or  C*Q**H,  or  C*Q,  where  Q is a complex unitary
                 matrix defined as the product of K elementary reflectors

       cunmql - overwrite the general complex M-by-N matrix  C  with  Q*C,  or
                 Q**H*C,  or  C*Q**H,  or  C*Q,  where  Q is a complex unitary
                 matrix defined as the product of K elementary reflectors

       cunmqr - overwrite the general complex M-by-N matrix C  with    SIDE  =
                 'L' SIDE = 'R' TRANS = 'N'

       cunmr2 - multiply a general matrix by the unitary matrix from a RQ fac-
                 torization determined by cgerqf (unblocked algorithm)

       cunmr3 - multiply a general matrix by the unitary matrix from a RZ fac-
                 torization determined by ctzrzf (unblocked algorithm)

       cunmrq  -  overwrite  the  general complex M-by-N matrix C with Q*C, or
                 Q**H*C, or C*Q**H, or C*Q,  where  Q  is  a  complex  unitary
                 matrix defined as the product of K elementary reflectors

       cunmrz  -  overwrite  the  general complex M-by-N matrix C with  Q*C or
                 Q**H*C or C*Q**H or C*Q.

       cunmtr - overwrite the general complex M-by-N matrix  C  with  Q*C,  or
                 Q**H*C,  or C*Q**H, or C*Q, where Q is defined as the product
                 of elementary reflectors, as returned by CHETRD

       cupgtr - generate a complex unitary matrix Q which is  defined  as  the
                 product  of  n-1  elementary  reflectors  H(i) of order n, as
                 returned by CHPTRD using packed storage

       cupmtr - overwrite the general complex M-by-N matrix C  with    SIDE  =
                 'L' SIDE = 'R' TRANS = 'N'

       cvbrmm - variable block sparse row format matrix-matrix multiply

       cvbrsm - variable block sparse row format triangular solve

       cvmul - compute the scaled product of complex vectors

       dCopy_CompCol_Matrix  -  A  utility  C  function  in the serial SuperLU
                 solver that copies one SuperMatrix into another.

       dCreate_CompCol_Matrix - A utility C function  in  the  serial  SuperLU
                 solver that creates a SuperMatrix in compressed sparse column
                 format (also known as the Harwell-Boeing format).

       dCreate_CompRow_Matrix - A utility C function  in  the  serial  SuperLU
                 solver  that  creates  a SuperMatrix in compressed sparse row
                 format.

       dCreate_Dense_Matrix - A utility  C  function  in  the  serial  SuperLU
                 solver that creates a SuperMatrix in dense format.

       dCreate_SuperNode_Matrix  -  A utility C function in the serial SuperLU
                 solver that creates a SuperMatrix in supernodal format.

       dPrintPerf - A utility function of the SuperLU solver that prints  sta-
                 tistics collected by the computational routines.

       dQuerySpace  - A inquiry function that provides information on the mem-
                 ory statistics of the SuperLU solver.

       dasum - Return the sum of the absolute values of a vector x.

       daxpy - compute y := alpha * x + y

       daxpyi - Compute y := alpha * x + y

       dbbcsd - compute the CS decomposition of an orthogonal matrix in  bidi-
                 agonal-block form

       dbcomm - block coordinate matrix-matrix multiply

       dbdimm - block diagonal format matrix-matrix multiply

       dbdism -  block diagonal format triangular solve

       dbdsdc - compute the singular value decomposition (SVD) of a real N-by-
                 N (upper or lower) bidiagonal matrix B

       dbdsqr - compute the singular value decomposition (SVD) of a real N-by-
                 N (upper or lower) bidiagonal matrix B

       dbelmm - block Ellpack format matrix-matrix multiply

       dbelsm - block Ellpack format triangular solve

       dbscmm - block sparse column matrix-matrix multiply

       dbscsm - block sparse column format triangular solve

       dbsrmm - block sparse row format matrix-matrix multiply

       dbsrsm - block sparse row format triangular solve

       dcnvcor - compute the convolution or correlation of real vectors

       dcnvcor2 - compute the convolution or correlation of real matrices

       dcoomm - coordinate matrix-matrix multiply

       dcopy - copy x to y

       dcosqb - synthesize a Fourier sequence from its representation in terms
                 of a cosine series with odd wave numbers. The COSQ operations
                 are  unnormalized  inverses of themselves, so a call to COSQF
                 followed by a call to COSQB will multiply the input  sequence
                 by 4 * N.

       dcosqf  - compute the Fourier coefficients in a cosine series represen-
                 tation with only odd wave numbers. The  COSQ  operations  are
                 unnormalized  inverses of themselves, so a call to COSQF fol-
                 lowed by a call to COSQB will multiply the input sequence  by
                 4 * N.

       dcosqi  -  initialize  the array WSAVE, which is used in both COSQF and
                 COSQB.

       dcost - compute the  discrete  Fourier  cosine  transform  of  an  even
                 sequence.   The  COST transforms are unnormalized inverses of
                 themselves, so a call of COST followed  by  another  call  of
                 COST will multiply the input sequence by 2 * (N-1).

       dcosti - initialize the array WSAVE, which is used in COST.

       dcscmm - compressed sparse column format matrix-matrix multiply

       dcscsm - compressed sparse column format triangular solve

       dcsrmm - compressed sparse row format matrix-matrix multiply

       dcsrsm - compressed sparse row format triangular solve

       ddiamm - diagonal format matrix-matrix multiply

       ddiasm - diagonal format triangular solve

       ddisna  - compute the reciprocal condition numbers for the eigenvectors
                 of a real symmetric or complex Hermitian matrix  or  for  the
                 left or right singular vectors of a general m-by-n matrix

       ddot - compute the dot product of two vectors x and y.

       ddoti - Compute the indexed dot product.

       dellmm - Ellpack format matrix-matrix multiply

       dellsm - Ellpack format triangular solve

       dezftb  -  computes  a periodic sequence from its Fourier coefficients.
                 DEZFTB is a simplified but slower version of DFFTB.

       dezftf - computes the Fourier  coefficients  of  a  periodic  sequence.
                 DEZFTF is a simplified but slower version of DFFTF.

       dezfti  - initializes the array WSAVE, which is used in both DEZFTF and
                 DEZFTB.

       dfft2b - compute a periodic sequence  from  its  Fourier  coefficients.
                 The  DFFT  operations  are  unnormalized, so a call of DFFT2F
                 followed by a call of DFFT2B will multiply the input sequence
                 by M*N.

       dfft2f  - compute the Fourier coefficients of a periodic sequence.  The
                 DFFT operations are unnormalized, so a call  of  DFFT2F  fol-
                 lowed by a call of DFFT2B will multiply the input sequence by
                 M*N.

       dfft2i - initialize the array WSAVE, which is used in both the  forward
                 and backward transforms.

       dfft3b - compute a periodic sequence from its Fourier coefficients. The
                 DFFT operations are unnormalized, so a call  of  DFFT3F  fol-
                 lowed by a call of DFFT3B will multiply the input sequence by
                 M*N*K.

       dfft3f - compute the Fourier coefficients of a real periodic  sequence.
                 The  DFFT  operations  are  unnormalized, so a call of DFFT3F
                 followed by a call of DFFT3B will multiply the input sequence
                 by M*N*K.

       dfft3i  -  initialize the array WSAVE, which is used in both DFFT3F and
                 DFFT3B.

       dfftb - compute a periodic sequence from its Fourier coefficients.  The
                 DFFT operations are unnormalized, so a call of DFFTF followed
                 by a call of DFFTB will multiply the input sequence by N.

       dfftf - compute the Fourier coefficients of a periodic  sequence.   The
                 FFT  operations are unnormalized, so a call of DFFTF followed
                 by a call of DFFTB will multiply the input sequence by N.

       dffti - initialize the array WSAVE, which is used  in  both  DFFTF  and
                 DFFTB.

       dfftopt - compute the length of the closest fast FFT

       dfftz  -  initialize the trigonometric weight and factor tables or com-
                 pute the forward Fast Fourier Transform of a double precision
                 sequence.

       dfftz2  - initialize the trigonometric weight and factor tables or com-
                 pute the two-dimensional forward Fast Fourier Transform of  a
                 two-dimensional double precision array.

       dfftz3  - initialize the trigonometric weight and factor tables or com-
                 pute the three-dimensional forward Fast Fourier Transform  of
                 a three-dimensional double complex array.

       dfftzm  - initialize the trigonometric weight and factor tables or com-
                 pute the one-dimensional forward Fast Fourier Transform of  a
                 set of double precision data sequences stored in a two-dimen-
                 sional array.

       dgbbrd - reduce a real general m-by-n band matrix A to upper bidiagonal
                 form B by an orthogonal transformation

       dgbcon - estimate the reciprocal of the condition number of a real gen-
                 eral band matrix A, in either the  1-norm  or  the  infinity-
                 norm, using the LU factorization computed by DGBTRF

       dgbequ  - compute row and column scalings intended to equilibrate an M-
                 by-N band matrix A and reduce its condition number

       dgbequb - compute row and column scalings intended to equilibrate an M-
                 by-N matrix A and reduce its condition number

       dgbmv  -  perform  one  of the matrix-vector operationsy := alpha*A*x +
                 beta*y or y := alpha*A'*x + beta*y

       dgbrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix  is  banded, and provide error
                 bounds and backward error estimates for the solution

       dgbrfsx - improve the computed solution to a system of linear equations
                 and provide error bounds and backward error estimates for the
                 solution

       dgbsv - compute the solution to  a  real  system  of  linear  equations
                 A*X=B,  where A is a band matrix of order N with KL subdiago-
                 nals and KU superdiagonals, and X and B are N-by-NRHS  matri-
                 ces

       dgbsvx  -  use  the  LU factorization to compute the solution to a real
                 system of linear  equations  A*X=B,  A**T*X=B,  or  A**H*X=B,
                 where A is a band matrix

       dgbsvxx  - compute the solution to system of linear equations A * X = B
                 for general band matrices

       dgbtf2 - compute an LU factorization of a real  m-by-n  band  matrix  A
                 using partial pivoting with row interchanges

       dgbtrf  -  compute  an  LU factorization of a real m-by-n band matrix A
                 using partial pivoting with row interchanges

       dgbtrs - solve a system of linear equations A*X=B or A'*X=B with a gen-
                 eral  band  matrix  A  using the LU factorization computed by
                 DGBTRF

       dgebak - form the right or left eigenvectors of a real  general  matrix
                 by  backward  transformation  on the computed eigenvectors of
                 the balanced matrix output by DGEBAL

       dgebal - balance a general real matrix A

       dgebd2 - reduce a general matrix to bidiagonal form using an  unblocked
                 algorithm

       dgebrd  - reduce a general real M-by-N matrix A to upper or lower bidi-
                 agonal form B by an orthogonal transformation

       dgecon - estimate the reciprocal of the condition number of  a  general
                 real  matrix  A,  in  either the 1-norm or the infinity-norm,
                 using the LU factorization computed by DGETRF

       dgeequ - compute row and column scalings intended to equilibrate an  M-
                 by-N matrix A and reduce its condition number

       dgeequb - compute row and column scalings intended to equilibrate an M-
                 by-N matrix A and reduce its condition number

       dgees - compute for an N-by-N real nonsymmetric matrix A, the eigenval-
                 ues,  the  real  Schur form T, and, optionally, the matrix of
                 Schur vectors Z

       dgeesx - compute for an N-by-N real nonsymmetric matrix A,  the  eigen-
                 values, the real Schur form T, and, optionally, the matrix of
                 Schur vectors Z

       dgeev - compute for an N-by-N real nonsymmetric matrix A, the eigenval-
                 ues and, optionally, the left and/or right eigenvectors

       dgeevx  -  compute for an N-by-N real nonsymmetric matrix A, the eigen-
                 values and, optionally, the left and/or right eigenvectors

       dgegs - routine is deprecated and has been replaced by routine DGGES

       dgegv - routine is deprecated and has been replaced by routine DGGEV

       dgehd2 - reduce a general square matrix to upper Hessenberg form  using
                 an unblocked algorithm

       dgehrd  -  reduce a real general matrix A to upper Hessenberg form H by
                 an orthogonal similarity transformation

       dgejsv - compute the singular value decomposition (SVD) of a real M-by-
                 N matrix A, where M >= N

       dgelq2  -  compute the LQ factorization of a general rectangular matrix
                 using an unblocked algorithm

       dgelqf - compute an LQ factorization of a real M-by-N matrix A

       dgels - solve overdetermined or  underdetermined  real  linear  systems
                 involving an M-by-N matrix A, or its transpose, using a QR or
                 LQ factorization of A

       dgelsd - compute the minimum-norm  solution  to  a  real  linear  least
                 squares problem

       dgelss  -  compute  the  minimum  norm  solution to a real linear least
                 squares problem

       dgelsx - routine is deprecated and has been replaced by routine DGELSY

       dgelsy - compute the minimum-norm  solution  to  a  real  linear  least
                 squares problem

       dgemm  -  perform  one  of the matrix-matrix operationsC := alpha*op( A
                 )*op( B ) + beta*C

       dgemqrt - overwrite the general real M-by-N matrix  C  with  Q*C,  C*Q,
                 Q**T*C, or C*Q**T depending on values of SIDE and TRANS

       dgemv  -  perform  one  of the matrix-vector operationsy := alpha*A*x +
                 beta*y or y := alpha*A'*x + beta*y

       dgeql2 - compute the QL factorization of a general  rectangular  matrix
                 using an unblocked algorithm

       dgeqlf - compute a QL factorization of a real M-by-N matrix A

       dgeqp3 - compute a QR factorization with column pivoting of a matrix A

       dgeqpf - routine is deprecated and has been replaced by routine DGEQP3

       dgeqr2  - computes the QR factorization of a general rectangular matrix
                 using an unblocked algorithm.

       dgeqr2p - computes the QR factorization of a general rectangular matrix
                 with  non-negative diagonal elements using an unblocked algo-
                 rithm.

       dgeqrf - compute a QR factorization of a real M-by-N matrix A

       dgeqrfp - compute a QR factorization of a real M-by-N matrix A: A = Q *
                 R

       dgeqrt  -  compute a blocked QR factorization of a real M-by-N matrix A
                 using the compact WY representation of Q

       dgeqrt2 - compute a QR factorization of a general real matrix using the
                 compact WY representation of Q

       dgeqrt3  -  recursively  compute  a  QR factorization of a general real
                 matrix using the compact WY representation of Q

       dger - perform the rank 1 operation A := alpha*x*y' + A

       dgerfs - improve the computed solution to a system of linear  equations
                 and provide error bounds and backward error estimates for the
                 solution

       dgerfsx - improve the computed solution to a system of linear equations
                 and provide error bounds and backward error estimates for the
                 solution

       dgerq2 - computes the RQ factorization of a general rectangular  matrix
                 using an unblocked algorithm

       dgerqf - compute an RQ factorization of a real M-by-N matrix A

       dgesdd - compute the singular value decomposition (SVD) of a real M-by-
                 N matrix A, optionally computing the left and right  singular
                 vectors

       dgesv  -  compute  the  solution  to  a real system of linear equations
                 A*X=B, where A is an N-by-N matrix and X and B are  N-by-NRHS
                 matrices

       dgesvd - compute the singular value decomposition (SVD) of a real M-by-
                 N matrix A, optionally computing the left and/or right singu-
                 lar vectors

       dgesvj - compute the singular value decomposition (SVD) of a real M-by-
                 N matrix A, where M >= N

       dgesvx - use the LU factorization to compute the  solution  to  a  real
                 system  of linear equations  A*X=B, where A is an N-by-N gen-
                 eral matrix

       dgesvxx - compute the solution to system of linear equations A*X=B  for
                 geberal matrices

       dgetf2 - compute an LU factorization of a general m-by-n matrix A using
                 partial pivoting with row interchanges

       dgetrf - compute an LU factorization of a general M-by-N matrix A using
                 partial pivoting with row interchanges

       dgetri  -  compute  the  inverse of a matrix using the LU factorization
                 computed by DGETRF

       dgetrs - solve a system of linear equations  A * X = B or A' *  X  =  B
                 with  a  general  N-by-N  matrix A using the LU factorization
                 computed by DGETRF

       dggbak - form the right or left eigenvectors of a real generalized  ei-
                 genvalue problem A*x = lambda*B*x, by backward transformation
                 on the computed eigenvectors of the balanced pair of matrices
                 output by DGGBAL

       dggbal - balance a pair of general real matrices (A,B)

       dgges - compute for a pair of N-by-N real nonsymmetric matrices (A,B),

       dggesx - compute for a pair of N-by-N real nonsymmetric matrices (A,B),
                 the generalized eigenvalues, the real Schur form (S,T),  and,
                 optionally, the left and/or right matrices of Schur vectors

       dggev - compute for a pair of N-by-N real nonsymmetric matrices (A,B)

       dggevx - compute for a pair of N-by-N real nonsymmetric matrices (A,B),
                 and, optionally, the left and/or right generalized  eigenvec-
                 tors

       dggglm - solve a general Gauss-Markov linear model (GLM) problem

       dgghrd - reduce a pair of real matrices (A,B) to generalized upper Hes-
                 senberg form using orthogonal transformations, where A  is  a
                 general matrix and B is upper triangular

       dgglse  -  solve  the  linear  equality-constrained least squares (LSE)
                 problem

       dggqrf - compute a generalized QR factorization of an N-by-M  matrix  A
                 and an N-by-P matrix B.

       dggrqf  -  compute a generalized RQ factorization of an M-by-N matrix A
                 and a P-by-N matrix B

       dggsvd - compute the generalized singular value decomposition (GSVD) of
                 an M-by-N real matrix A and P-by-N real matrix B

       dggsvp - compute orthogonal matrices

       dgscon  - estimates the reciprocal of the condition number of a general
                 real matrix A, in either the  1-norm  or  the  infinity-norm,
                 using  the  LU  factorization  computed  by  SuperLU  routine
                 sgstrf.

       dgsequ - computes row and column scalings intended to equilibrate an M-
                 by-N sparse matrix A and reduce its condition number.

       dgsrfs - improves the computed solution to a system of linear equations
                 and provides error bounds and backward  error  estimates  for
                 the solution.  It is a SuperLU routine.

       dgssco - General sparse solver condition number estimate.

       dgssda - Deallocate working storage for the general sparse solver.

       dgssfa - General sparse solver numeric factorization.

       dgssfs - General sparse solver one call interface.

       dgssin - Initialize the general sparse solver.

       dgssor - General sparse solver ordering and symbolic factorization.

       dgssps - Print general sparse solver statics.

       dgssrp - Return permutation used by the general sparse solver.

       dgsssl - Solve routine for the general sparse solver.

       dgssuo - Provide general sparse solvers SPSOLVE and SuperLU a user-sup-
                 plied permutation for ordering.

       dgssv - solves a system of linear equations A*X=B using the LU  factor-
                 ization from sgstrf.

       dgssvx  -  solves the system of linear equations A*X=B or A'*X=B, using
                 the LU factorization from sgstrf(). Error bounds on the solu-
                 tion and a condition estimate are also provided.

       dgstrf - computes an LU factorization of a general sparse m-by-n matrix
                 A using partial pivoting with row interchanges.

       dgstrs - solves a system of linear equations A*X=B  or  A'*X=B  with  A
                 sparse  and  B  dense, using the LU factorization computed by
                 sgstrf.

       dgsvj0 - pre-processor for the routine sgesvj

       dgsvj1 - pre-processor for the routine sgesvj, apply  Jacobi  rotations
                 targeting only particular pivots

       dgtcon  -  estimate  the  reciprocal  of the condition number of a real
                 tridiagonal matrix A using the LU factorization  as  computed
                 by DGTTRF

       dgthr - Gathers specified elements from y into x.

       dgthrz - Gather and zero.

       dgtrfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix  is  tridiagonal,  provide  error
                 bounds and backward error estimates for the solution

       dgtsv  -  solve  the  equation  A*X=B, where A is an N-by-N tridiagonal
                 matrix, by Gaussian elimination with partial pivoting

       dgtsvx - use the LU factorization to compute the  solution  to  a  real
                 system  of  linear  equations A*X=B or A**T*X=B, where A is a
                 tridiagonal matrix of order N  and  X  and  B  are  N-by-NRHS
                 matrices

       dgttrf  -  compute  an  LU factorization of a real tridiagonal matrix A
                 using elimination with partial pivoting and row interchanges

       dgttrs - solve one of the systems of equations A*X=B or A'*X=B, with  a
                 tridiagonal  matrix  A using the LU factorization computed by
                 DGTTRF

       dgtts2 - solve a system of linear equations with a  tridiagonal  matrix
                 using the LU factorization computed by dgttrf

       dhgeqz  - implement a single-/double-shift version of the QZ method for
                 finding  the  generalized  eigenvalues    w(j)=(ALPHAR(j)   +
                 i*ALPHAI(j))/BETAR(j)  of  the equation   det( A-w(i) B ) = 0
                 In addition, the pair A,B may be reduced to generalized Schur
                 form

       dhsein  -  use  inverse  iteration  to find specified right and/or left
                 eigenvectors of a real upper Hessenberg matrix H

       dhseqr - compute the eigenvalues of a real upper  Hessenberg  matrix  H
                 and, optionally, the matrices T and Z from the Schur decompo-
                 sition H = Z T Z**T, where T  is  an  upper  quasi-triangular
                 matrix  (the  Schur  form), and Z is the orthogonal matrix of
                 Schur vectors

       dinfinite_norm_error - A utility function of the  SuperLU  solver  that
                 computes  the  infinity-norm  of an array of vectors that are
                 approximations to the exact solution vector.

       djadmm - Jagged diagonal matrix-matrix multiply (modified Ellpack)

       djadrp - right permutation of a jagged diagonal matrix

       djadsm - Jagged-diagonal format triangular solve

       dla_gbamv - perform a matrix-vector operation to calculate error bounds

       dla_gbrcond  - estimate the Skeel condition number for a general banded
                 matrix

       dla_gbrfsx_extended - improve the computed solution to a system of lin-
                 ear  equations  for  general  banded  matrices  by performing
                 extra-precise iterative refinement and provide  error  bounds
                 and backward error estimates for the solution

       dla_gbrpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U) for a general banded matrix

       dla_geamv - compute a matrix-vector product using a general  matrix  to
                 calculate error bounds

       dla_gercond - estimate the Skeel condition number for a general matrix

       dla_gerfsx_extended - improve the computed solution to a system of lin-
                 ear equations for general matrices by  performing  extra-pre-
                 cise  iterative refinement and provide error bounds and back-
                 ward error estimates for the solution

       dla_gerpvgrw - compute the reciprocal pivot  growth  factor  using  the
                 "max absolute element" norm

       dla_lin_berr - compute a component-wise relative backward error

       dla_porcond - estimate the Skeel condition number for a symmetric posi-
                 tive-definite matrix

       dla_porfsx_extended - improve the computed solution to a system of lin-
                 ear  equations  for  symmetric or Hermitian positive-definite
                 matrices by performing extra-precise iterative refinement and
                 provide  error  bounds  and  backward error estimates for the
                 solution

       dla_porpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U)  for  a symmetric or Hermitian positive-defi-
                 nite matrix

       dla_syamv - compute a matrix-vector product using a  symmetric  indefi-
                 nite matrix to calculate error bounds

       dla_syrcond  -  estimate  the  Skeel  condition  number for a symmetric
                 indefinite matrix

       dla_syrfsx_extended - improve the computed solution to a system of lin-
                 ear equations for symmetric indefinite matrices by performing
                 extra-precise iterative refinement and provide  error  bounds
                 and backward error estimates for the solution

       dla_syrpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U) for a symmetric indefinite matrix

       dla_wwaddw - add a vector into a doubled-single vector

       dlaed0 - compute all eigenvalues and corresponding eigenvectors  of  an
                 unreduced  symmetric  tridiagonal matrix using the divide and
                 conquer method. Used by dstedc

       dlaed1 - compute the updated eigensystem of  a  diagonal  matrix  after
                 modification  by a rank-one symmetric matrix. Used by dstedc,
                 when the original matrix is tridiagonal

       dlaed2 - merge eigenvalues and deflates secular equation; used by  dst-
                 edc when the original matrix is tridiagonal

       dlaed3  - find the roots of the secular equation and updates the eigen-
                 vectors; used by dstedc when the original matrix is tridiago-
                 nal

       dlaed4 - is used by sstedc. Finds a single root of the secular equation

       dlaed5 - is used by sstedc. Solve the 2-by-2 secular equation

       dlaed6 - is used by sstedc. Compute one Newton step in solution of  the
                 secular equation

       dlaed7  -  compute  the  updated eigensystem of a diagonal matrix after
                 modification by a rank-one symmetric matrix. Used by  dstedc,
                 when the original matrix is dense

       dlaed8  - merge eigenvalues and deflates secular equation. Used by dst-
                 edc, when the original matrix is dense

       dlaed9 - is used by sstedc. Find the roots of the secular equation  and
                 updates  the  eigenvectors.  Used when the original matrix is
                 dense

       dlaeda - is used by sstedc. Compute the Z vector determining the  rank-
                 one modification of the diagonal matrix. Used when the origi-
                 nal matrix is dense

       dlagtf - factorize the matrix (T-lambda*I), where T is an n by n tridi-
                 agonal matrix and lambda is a scalar, as T-lambda*I = PLU

       dlals0  -  apply  back multiplying factors in solving the least squares
                 problem using divide and conquer SVD approach. Used by dgelsd

       dlalsa  -  compute  the  SVD of the coefficient matrix in compact form.
                 Used by dgelsd

       dlalsd - use the singular value decomposition of A to solve  the  least
                 squares problem

       dlamch - Determines double precision machine parameters.

       dlamrg  -  will create a permutation list which will merge the elements
                 of A (which is composed of  two  independently  sorted  sets)
                 into a single set which is sorted in ascending order

       dlangs  -  returns the value of the one-norm, or the Frobenius-norm, or
                 the infinity-norm, or the element with largest absolute value
                 of a general real matrix A in sparse format.

       dlansf  - return the value of the 1-norm, or the Frobenius norm, or the
                 infinity norm, or the element of largest absolute value of  a
                 symmetric matrix in RFP format

       dlaqgs  -  a SuperLU function that equilibrates a general sparse M by N
                 matrix A.

       dlarscl2 - perform reciprocal diagonal scaling on a vector

       dlartgs - generate a plane rotation designed to introduce  a  bulge  in
                 implicit QR iteration for the bidiagonal SVD problem

       dlarz  - apply a real elementary reflector H to a real M-by-N matrix C,
                 from either the left or the right

       dlarzb - apply a real block reflector H or its transpose H**T to a real
                 distributed M-by-N C from the left or the right

       dlarzt  -  form  the triangular factor T of a real block reflector H of
                 order > n, which is defined as  a  product  of  k  elementary
                 reflectors

       dlascl2 - perform diagonal scaling on a vector

       dlasq1  -  compute  the  singular  values  of  a real square bidiagonal
                 matrix. Used by sbdsqr

       dlasq2 -  compute all the eigenvalues of a real symmetric positive def-
                 inite tridiagonal matrix (high relative accuracy)

       dlasq3  - check for deflation, computes a shift and calls dqds. Used by
                 dbdsqr

       dlasq4 - compute an approximation to the smallest eigenvalue using val-
                 ues of d from the previous transform. Used by dbdsqr

       dlasq5  -  compute one dqds transform in ping-pong form. Used by sbdsqr
                 and sstegr

       dlasq6 - compute one dqd transform in ping-pong form.  Used  by  sbdsqr
                 and sstegr

       dlasrt  -  the  numbers  in  D  in increasing order (if ID = 'I') or in
                 decreasing order (if ID = 'D' )

       dlasyf - compute a partial factorization of  a  real  symmetric  matrix
                 using the Bunch-Kaufman diagonal pivoting method

       dlasyf_rook  -  compute  a  partial  factorization  of a real symmetric
                 matrix using the bounded Bunch-Kaufman ("rook") diagonal piv-
                 oting method

       dlatrz  -  factor  an  upper  trapezoidal matrix by means of orthogonal
                 transformations

       dlatzm - routine is deprecated and has been replaced by routine DORMRZ

       dnrm2 - Return the Euclidian norm of a vector.

       dopgtr - generate a real orthogonal matrix Q which is  defined  as  the
                 product  of  n-1  elementary  reflectors  H(i) of order n, as
                 returned by DSPTRD using packed storage

       dopmtr - overwrite the general real M-by-N matrix C with   SIDE  =  'L'
                 SIDE = 'R' TRANS = 'N'

       dorbdb  -  simultaneously  bidiagonalize the blocks of an M-by-M parti-
                 tioned orthogonal matrix

       dorbdb1 - simultaneously bidiagonalize the blocks of a tall and  skinny
                 matrix with orthonomal columns

       dorbdb2  - simultaneously bidiagonalize the blocks of a tall and skinny
                 matrix with orthonomal columns

       dorbdb3 - simultaneously bidiagonalize the blocks of a tall and  skinny
                 matrix with orthonomal columns

       dorbdb4  - simultaneously bidiagonalize the blocks of a tall and skinny
                 matrix with orthonomal columns

       dorbdb5 - orthogonalize one column vector with respect to the orthonor-
                 mal columns of another column vector

       dorbdb6 - orthogonalize one column vector with respect to the orthonor-
                 mal columns of another column vector

       dorcsd - compute the CS decomposition of an M-by-M partitioned orthogo-
                 nal matrix

       dorcsd2by1  -  compute  the  CS  decomposition of an M-by-Q matrix with
                 orthonormal columns that has been partitioned into  a  2-by-1
                 block structure

       dorg2l - generate an m by n real matrix Q with orthonormal columns,

       dorg2r - generate an m by n real matrix Q with orthonormal columns,

       dorgbr  - generate one of the real orthogonal matrices Q or P**T deter-
                 mined by DGEBRD when reducing a real matrix A  to  bidiagonal
                 form

       dorghr  -  generate  a real orthogonal matrix Q which is defined as the
                 product of IHI-ILO  elementary  reflectors  of  order  N,  as
                 returned by DGEHRD

       dorgl2 - generate an m by n real matrix Q with orthonormal rows,

       dorglq - generate an M-by-N real matrix Q with orthonormal rows,

       dorgql - generate an M-by-N real matrix Q with orthonormal columns,

       dorgqr - generate an M-by-N real matrix Q with orthonormal columns,

       dorgr2 - generate an m by n real matrix Q with orthonormal rows,

       dorgrq - generate an M-by-N real matrix Q with orthonormal rows,

       dorgtr  -  generate  a real orthogonal matrix Q which is defined as the
                 product of n-1 elementary reflectors of order N, as  returned
                 by DSYTRD

       dorm2l  -  multiply a general matrix by the orthogonal matrix from a QL
                 factorization determined by sgeqlf (unblocked algorithm)

       dorm2r - multiply a general matrix by the orthogonal matrix from  a  QR
                 factorization determined by sgeqrf (unblocked algorithm)

       dormbr - overwrites the general real M-by-N matrix C with Q*C or Q**T*C
                 or C*Q**T or C*Q or P*C or P**T*C or C*P or C*P**T.

       dormhr - overwrite the general real M-by-N matrix C with Q*C or  Q**T*C
                 or C*Q**T or C*Q.

       dorml2  -  multiply a general matrix by the orthogonal matrix from a LQ
                 factorization determined by sgelqf (unblocked algorithm)

       dormlq - overwrite the general real M-by-N matrix C with Q*C or  Q**T*C
                 or C*Q**T or C*Q.

       dormql  - overwrite the general real M-by-N matrix C with Q*C or Q**T*C
                 or C*Q**T or C*Q.

       dormqr - overwrite the general real M-by-N matrix C with   SIDE  =  'L'
                 SIDE = 'R' TRANS = 'N'

       dormr2  -  multiply a general matrix by the orthogonal matrix from a RQ
                 factorization determined by sgerqf (unblocked algorithm)

       dormr3 - multiply a general matrix by the orthogonal matrix from  a  RZ
                 factorization determined by stzrzf (unblocked algorithm)

       dormrq  - overwrite the general real M-by-N matrix C with Q*C or Q**T*C
                 or C*Q**T or C*Q.

       dormrz - overwrite the general real M-by-N matrix C with Q*C or  Q**H*C
                 or C*Q**H or C*Q.

       dormtr  - overwrite the general real M-by-N matrix C with Q*C or Q**T*C
                 or C*Q**T or C*Q.

       dpbcon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm)  of  a  real  symmetric positive definite band matrix
                 using the Cholesky factorization A = U**T*U  or  A  =  L*L**T
                 computed by DPBTRF

       dpbequ - compute row and column scalings intended to equilibrate a sym-
                 metric positive definite band matrix A and reduce its  condi-
                 tion number (with respect to the two-norm)

       dpbrfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is  symmetric  positive  definite
                 and banded, provide error bounds and backward error estimates
                 for the solution

       dpbstf - compute a split Cholesky factorization  of  a  real  symmetric
                 positive definite band matrix A

       dpbsv  -  compute  the  solution  to  a real system of linear equations
                 A*X=B, where A is an N-by-N symmetric positive definite  band
                 matrix and X and B are N-by-NRHS matrices

       dpbsvx  -  use  the Cholesky factorization to compute the solution to a
                 real system of linear equations A*X=B, where A is  an  N-by-N
                 symmetric positive definite band matrix and X and B are N-by-
                 NRHS matrices

       dpbtf2 - compute the Cholesky factorization of a real  symmetric  posi-
                 tive definite band matrix A

       dpbtrf  -  compute the Cholesky factorization of a real symmetric posi-
                 tive definite band matrix A

       dpbtrs - solve a system of linear equations A*X = B  with  a  symmetric
                 positive definite band matrix A using the Cholesky factoriza-
                 tion A = U**T*U or A = L*L**T computed by DPBTRF

       dpftrf - compute the Cholesky factorization of a real  symmetric  posi-
                 tive definite matrix A, the block version of the algorithm

       dpftri  -  compute  the  inverse  of a real symmetric positive definite
                 matrix A using the Cholesky factorization computed by DPFTRF

       dpftrs - solve a system of linear equations A*X = B  with  a  symmetric
                 positive  definite  matrix A using the Cholesky factorization
                 computed by DPFTRF

       dpocon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm)  of  a  real symmetric positive definite matrix using
                 the Cholesky factorization A = U**T*U or A = L*L**T  computed
                 by DPOTRF

       dpoequ - compute row and column scalings intended to equilibrate a sym-
                 metric positive definite matrix A and  reduce  its  condition
                 number (with respect to the two-norm)

       dpoequb  -  compute  row  and column scalings intended to equilibrate a
                 symmetric positive definite matrix A and reduce its condition
                 number with respect to the two-norm

       dporfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is symmetric  positive  definite,
                 provide  error  bounds  and  backward error estimates for the
                 solution

       dporfsx - improve the computed solution to a system of linear equations
                 when  the  coefficient matrix is symmetric positive definite,
                 provide error bounds and backward  error  estimates  for  the
                 solution

       dposv  - compute the solution to a real system of linear equations  A*X
                 = B, where A is an N-by-N symmetric positive definite  matrix
                 and X and B are N-by-NRHS matrices

       dposvx  -  use  the Cholesky factorization to compute the solution to a
                 real system of linear equations  A*X = B, where A is an N-by-
                 N  symmetric  positive  definite matrix and X and B are N-by-
                 NRHS matrices

       dposvxx - compute the solution to a double precision system  of  linear
                 equations  A*X  =  B, where A is an N-by-N symmetric positive
                 definite matrix and X and B are N-by-NRHS matrices

       dpotf2 - compute the Cholesky factorization of a real  symmetric  posi-
                 tive definite matrix A

       dpotrf  -  compute the Cholesky factorization of a real symmetric posi-
                 tive definite matrix A

       dpotri - compute the inverse of  a  real  symmetric  positive  definite
                 matrix  A  using the Cholesky factorization A = U**T*U or A =
                 L*L**T computed by DPOTRF

       dpotrs - solve a system of linear equations A*X = B  with  a  symmetric
                 positive definite matrix A using the Cholesky factorization A
                 = U**T*U or A = L*L**T computed by DPOTRF

       dppcon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm)  of  a real symmetric positive definite packed matrix
                 using the Cholesky factorization A = U**T*U  or  A  =  L*L**T
                 computed by DPPTRF

       dppequ - compute row and column scalings intended to equilibrate a sym-
                 metric positive definite  matrix  A  in  packed  storage  and
                 reduce its condition number (with respect to the two-norm)

       dpprfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is  symmetric  positive  definite
                 and packed, provide error bounds and backward error estimates
                 for the solution

       dppsv - compute the solution to a real system of linear equations A*X =
                 B,  where  A  is an N-by-N symmetric positive definite matrix
                 stored in packed format and X and B are N-by-NRHS matrices

       dppsvx - use the Cholesky factorization to compute the  solution  to  a
                 real system of linear equations  A*X = B, where A is an N-by-
                 N symmetric positive definite matrix stored in packed  format
                 and X and B are N-by-NRHS matrices

       dpptrf  -  compute the Cholesky factorization of a real symmetric posi-
                 tive definite matrix A stored in packed format

       dpptri - compute the inverse of  a  real  symmetric  positive  definite
                 matrix  A  using the Cholesky factorization A = U**T*U or A =
                 L*L**T computed by DPPTRF

       dpptrs - solve a system of linear equations A*X = B  with  a  symmetric
                 positive  definite  matrix  A  in  packed  storage  using the
                 Cholesky factorization A = U**T*U or A = L*L**T  computed  by
                 DPPTRF

       dpstf2 - compute the Cholesky factorization with complete pivoting of a
                 real symmetric positive semidefinite matrix A

       dpstrf - compute the Cholesky factorization with complete pivoting of a
                 real symmetric positive semidefinite matrix A

       dptcon - compute the reciprocal of the condition number (in the 1-norm)
                 of a real  symmetric  positive  definite  tridiagonal  matrix
                 using the factorization A = L*D*L**T or A = U**T*D*U computed
                 by DPTTRF

       dpteqr - compute all eigenvalues and,  optionally,  eigenvectors  of  a
                 symmetric  positive definite tridiagonal matrix by first fac-
                 toring the matrix using DPTTRF, and then  calling  DBDSQR  to
                 compute the singular values of the bidiagonal factor

       dptrfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is  symmetric  positive  definite
                 and  tridiagonal,  provide  error  bounds  and backward error
                 estimates for the solution

       dptsv - compute the solution to a real system of linear equations A*X =
                 B, where A is an N-by-N symmetric positive definite tridiago-
                 nal matrix, and X and B are N-by-NRHS matrices

       dptsvx - use the factorization A = L*D*L**T to compute the solution  to
                 a  real  system of linear equations A*X = B, where A is an N-
                 by-N symmetric positive definite tridiagonal matrix and X and
                 B are N-by-NRHS matrices

       dpttrf  - compute the L*D*L' factorization of a real symmetric positive
                 definite tridiagonal matrix A

       dpttrs - solve a tridiagonal system of the form  A * X =  B  using  the
                 L*D*L' factorization of A computed by DPTTRF

       dptts2  -  solve  a tridiagonal system of the form  A * X = B using the
                 L*D*L' factorization of A computed by DPTTRF

       dqdota - compute a double precision constant plus an extended precision
                 constant  plus the extended precision dot product of two dou-
                 ble precision vectors x and y.

       dqdoti - compute a constant plus the extended precision dot product  of
                 two double precision vectors x and y.

       drot - Apply a Given's rotation constructed by SROTG.

       drotg - Construct a Given's plane rotation

       droti -  Apply an indexed Givens rotation.

       drotm  -  Apply  a Gentleman's modified Given's rotation constructed by
                 SROTMG.

       drotmg - Construct a Gentleman's modified Given's plane rotation

       dsbev - compute all the eigenvalues and, optionally, eigenvectors of  a
                 real symmetric band matrix A

       dsbevd - compute all the eigenvalues and, optionally, eigenvectors of a
                 real symmetric band matrix A

       dsbevx - compute selected eigenvalues and, optionally, eigenvectors  of
                 a real symmetric band matrix A

       dsbgst - reduce a real symmetric-definite banded generalized eigenprob-
                 lem A*x = lambda*B*x to standard form C*y = lambda*y,

       dsbgv - compute all the eigenvalues, and optionally,  the  eigenvectors
                 of a real generalized symmetric-definite banded eigenproblem,
                 of the form A*x=(lambda)*B*x

       dsbgvd - compute all the eigenvalues, and optionally, the  eigenvectors
                 of a real generalized symmetric-definite banded eigenproblem,
                 of the form A*x=(lambda)*B*x

       dsbgvx - compute selected eigenvalues, and optionally, eigenvectors  of
                 a real generalized symmetric-definite banded eigenproblem, of
                 the form A*x=(lambda)*B*x

       dsbmv - perform the matrix-vector operationy := alpha*A*x + beta*y

       dsbtrd - reduce a real symmetric band matrix A to symmetric tridiagonal
                 form T by an orthogonal similarity transformation

       dscal - Compute y := alpha * y

       dsctr - Scatters elements from x into y

       dsdot  -  compute the double precision dot product of two single preci-
                 sion vectors x and y.

       dsecnd - return the user time for a process in seconds

       dsfrk - perform a symmetric rank-k operation for matrix in RFP format

       dsgesv - compute the solution to a real system of linear equations  A*X
                 = B

       dsinqb - synthesize a Fourier sequence from its representation in terms
                 of a sine series with odd wave numbers.  The SINQ  operations
                 are  unnormalized  inverses of themselves, so a call to SINQF
                 followed by a call to SINQB will multiply the input  sequence
                 by 4 * N.

       dsinqf  - compute the Fourier coefficients in a sine series representa-
                 tion with only  odd  wave  numbers.The  SINQ  operations  are
                 unnormalized  inverses of themselves, so a call to SINQF fol-
                 lowed by a call to SINQB will multiply the input sequence  by
                 4 * N.

       dsinqi  -  initialize the array xWSAVE, which is used in both SINQF and
                 SINQB.

       dsint - compute the discrete Fourier sine transform of an odd sequence.
                 The  SINT transforms are unnormalized inverses of themselves,
                 so a call of SINT followed by another call of SINT will  mul-
                 tiply the input sequence by 2 * (N+1).

       dsinti  - initialize the array WSAVE, which is used in subroutine SINT.

       dskymm - Skyline format matrix-matrix multiply

       dskysm - Skyline format triangular solve

       dspcon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm) of a real symmetric packed matrix A using the factor-
                 ization A = U*D*U**T or A = L*D*L**T computed by DSPTRF

       dspev - compute all the eigenvalues and, optionally, eigenvectors of  a
                 real symmetric matrix A in packed storage

       dspevd - compute all the eigenvalues and, optionally, eigenvectors of a
                 real symmetric matrix A in packed storage

       dspevx - compute selected eigenvalues and, optionally, eigenvectors  of
                 a real symmetric matrix A in packed storage

       dspgst  -  reduce a real symmetric-definite generalized eigenproblem to
                 standard form, using packed storage

       dspgv - compute all the eigenvalues and, optionally,  the  eigenvectors
                 of a real generalized symmetric-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       dspgvd - compute all the eigenvalues, and optionally, the  eigenvectors
                 of a real generalized symmetric-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       dspgvx - compute selected eigenvalues, and optionally, eigenvectors  of
                 a  real  generalized  symmetric-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       dspmv - perform the matrix-vector operationy := alpha*A*x + beta*y

       dspr - perform the symmetric rank 1 operation   A := alpha*x*x' + A

       dspr2 - perform the symmetric rank 2  operation    A  :=  alpha*x*y'  +
                 alpha*y*x' + A

       dsprfs  - improve the computed solution to a system of linear equations
                 when the  coefficient  matrix  is  symmetric  indefinite  and
                 packed,  provides  error  bounds and backward error estimates
                 for the solution

       dspsv - compute the solution to a real system of linear equations   A*X
                 =  B,  where A is an N-by-N symmetric matrix stored in packed
                 format and X and B are N-by-NRHS matrices

       dspsvx - use the diagonal pivoting factorization A = U*D*U**T  or  A  =
                 L*D*L**T  to  compute the solution to a real system of linear
                 equations A * X = B, where A is an  N-by-N  symmetric  matrix
                 stored in packed format and X and B are N-by-NRHS matrices

       dsptrd - reduce a real symmetric matrix A stored in packed form to sym-
                 metric tridiagonal form T by an orthogonal similarity  trans-
                 formation

       dsptrf  - compute the factorization of a real symmetric matrix A stored
                 in packed format using the  Bunch-Kaufman  diagonal  pivoting
                 method

       dsptri - compute the inverse of a real symmetric indefinite matrix A in
                 packed storage using the factorization A = U*D*U**T  or  A  =
                 L*D*L**T computed by DSPTRF

       dsptrs - solve a system of linear equations A*X = B with a real symmet-
                 ric matrix A stored in packed format using the  factorization
                 A = U*D*U**T or A = L*D*L**T computed by DSPTRF

       dstebz - compute the eigenvalues of a symmetric tridiagonal matrix T

       dstedc  -  compute  all  eigenvalues and, optionally, eigenvectors of a
                 symmetric tridiagonal matrix using  the  divide  and  conquer
                 method

       dstegr - (a) Compute T-sigma_i = L_i D_i L_i^T, such that L_i D_i L_i^T
                 is a relatively robust representation

       dstein - compute the  eigenvectors  of  a  real  symmetric  tridiagonal
                 matrix   T  corresponding  to  specified  eigenvalues,  using
                 inverse iteration

       dstemr - compute selected eigenvalues and, optionally, eigenvectors  of
                 a real symmetric tridiagonal matrix T

       dsteqr  -  compute  all  eigenvalues and, optionally, eigenvectors of a
                 symmetric tridiagonal matrix using  the  implicit  QL  or  QR
                 method

       dsterf  -  compute  all  eigenvalues  of a symmetric tridiagonal matrix
                 using the Pal-Walker-Kahan variant of the QL or QR algorithm

       dstev - compute all eigenvalues and, optionally, eigenvectors of a real
                 symmetric tridiagonal matrix A

       dstevd  -  compute  all  eigenvalues and, optionally, eigenvectors of a
                 real symmetric tridiagonal matrix

       dstevr - compute selected eigenvalues and, optionally, eigenvectors  of
                 a real symmetric tridiagonal matrix T

       dstevx  - compute selected eigenvalues and, optionally, eigenvectors of
                 a real symmetric tridiagonal matrix A

       dstsv - compute the solution to a system of linear equations A * X =  B
                 where A is a symmetric tridiagonal matrix

       dsttrf  - compute the factorization of a symmetric tridiagonal matrix A
                 using the Bunch-Kaufman diagonal pivoting method

       dsttrs - compute the solution to a real system of linear equations  A*X
                 =  B, where A is an N-by-N symmetric tridiagonal matrix and X
                 and B are N-by-NRHS matrices

       dswap - Exchange vectors x and y

       dsycon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm)  of a real symmetric matrix A using the factorization
                 A = U*D*U**T or A = L*D*L**T computed by DSYTRF

       dsycon_rook - estimate the reciprocal of the condition number  (in  the
                 1-norm)  of a real symmetric matrix using the factorization A
                 = U*D*U**T or A = L*D*L**T computed by DSYTRF_ROOK

       dsyconv - convert A given by TRF into L and D and vice-versa

       dsyequb - compute row and column scalings  intended  to  equilibrate  a
                 symmetric  matrix  A  and  reduce  its  condition number with
                 respect to the two-norm

       dsyev - compute all eigenvalues and, optionally, eigenvectors of a real
                 symmetric matrix A

       dsyevd  -  compute  all  eigenvalues and, optionally, eigenvectors of a
                 real symmetric matrix A

       dsyevr - compute selected eigenvalues and, optionally, eigenvectors  of
                 a real symmetric tridiagonal matrix T

       dsyevx  - compute selected eigenvalues and, optionally, eigenvectors of
                 a real symmetric matrix A

       dsygs2 - reduce a real symmetric-definite generalized  eigenproblem  to
                 standard form

       dsygst  -  reduce a real symmetric-definite generalized eigenproblem to
                 standard form

       dsygv - compute all the eigenvalues, and optionally,  the  eigenvectors
                 of a real generalized symmetric-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       dsygvd - compute all the eigenvalues, and optionally, the  eigenvectors
                 of a real generalized symmetric-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       dsygvx - compute selected eigenvalues, and optionally, eigenvectors  of
                 a  real  generalized  symmetric-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       dsymm - perform one of the matrix-matrix  operationsC  :=  alpha*A*B  +
                 beta*C or C := alpha*B*A + beta*C

       dsymv - perform the matrix-vector operationy := alpha*A*x + beta*y

       dsyr - perform the symmetric rank 1 operation   A := alpha*x*x' + A

       dsyr2  -  perform  the  symmetric  rank 2 operation   A := alpha*x*y' +
                 alpha*y*x' + A

       dsyr2k - perform one  of  the  symmetric  rank  2k  operations    C  :=
                 alpha*A*B'  +  alpha*B*A'  +  beta*C  or  C  :=  alpha*A'*B +
                 alpha*B'*A + beta*C

       dsyrfs - improve the computed solution to a system of linear  equations
                 when  the coefficient matrix is symmetric indefinite, provide
                 error bounds and backward error estimates for the solution

       dsyrfsx - improve the computed solution to a system of linear equations
                 when  the coefficient matrix is symmetric indefinite, provide
                 error bounds and backward error estimates for the solution

       dsyrk -  perform  one  of  the  symmetric  rank  k  operations    C  :=
                 alpha*A*A' + beta*C or C := alpha*A'*A + beta*C

       dsysv  - compute the solution to a real system of linear equations  A*X
                 = B, where A is an N-by-N symmetric matrix and X and B are N-
                 by-NRHS matrices

       dsysv_rook - compute the solution to system of linear equations A*X = B
                 for symmetric matrices. DSYTRF_ROOK is called to compute  the
                 factorization of A

       dsysvx  -  use the diagonal pivoting factorization to compute the solu-
                 tion to a real system of linear equations A*X = B, where A is
                 an N-by-N symmetric matrix and X and B are N-by-NRHS matrices

       dsysvxx - compute the solution to real system of linear equations A*X =
                 B for symmetric matrices

       dsytd2 - reduce a real symmetric matrix A to symmetric tridiagonal form
                 T by an orthogonal similarity transformation

       dsytf2 - compute the  factorization  of  a  real  symmetric  indefinite
                 matrix,  using  the diagonal pivoting method (unblocked algo-
                 rithm)

       dsytf2_rook - compute the factorization of a real symmetric  indefinite
                 matrix using the bounded Bunch-Kaufman ("rook") diagonal piv-
                 oting method (unblocked algorithm)

       dsytrd - reduce a real symmetric matrix A to real symmetric tridiagonal
                 form T by an orthogonal similarity transformation

       dsytrf  -  compute the factorization of a real symmetric matrix A using
                 the Bunch-Kaufman diagonal pivoting method

       dsytrf_rook - compute the factorization  of  a  real  symmetric  matrix
                 using  the  bounded  Bunch-Kaufman ("rook") diagonal pivoting
                 method (blocked algorithm)

       dsytri - compute the inverse of a real symmetric  indefinite  matrix  A
                 using the factorization A = U*D*U**T or A = L*D*L**T computed
                 by DSYTRF

       dsytri2 - compute the inverse of a DOUBLE PRECISION  symmetric  indefi-
                 nite  matrix  A  using  the factorization A = U*D*U**T or A =
                 L*D*L**T computed by DSYTRF

       dsytri2x - compute the inverse of a real symmetric indefinite matrix  A
                 using the factorization computed by DSYTRF

       dsytri_rook - compute the inverse of a real symmetric indefinite matrix
                 A using the factorization A = U*D*U**T or A =  L*D*L**T  com-
                 puted by DSYTRF_ROOK

       dsytrs - solve a system of linear equations A*X = B with a real symmet-
                 ric matrix A using the factorization A  =  U*D*U**T  or  A  =
                 L*D*L**T computed by DSYTRF

       dsytrs2  -  solve a system of linear equations A*X = B with a real sym-
                 metric matrix A using the factorization  computed  by  DSYTRF
                 and converted by DSYCONV

       dsytrs_rook  -  solve  a system of linear equations A*X = B with a real
                 symmetric matrix A using the factorization A = U*D*U**T or  A
                 = L*D*L**T computed by DSYTRF_ROOK

       dtbcon  - estimate the reciprocal of the condition number of a triangu-
                 lar band matrix A, in either the 1-norm or the infinity-norm

       dtbmv - perform one of the matrix-vector operationsx := A*x,  or  x  :=
                 A'*x

       dtbrfs  -  provide  error  bounds  and backward error estimates for the
                 solution to a system of linear equations  with  a  triangular
                 band coefficient matrix

       dtbsv - solve one of the systems of equations A*x = b, or A'*x = b

       dtbtrs  -  solve a triangular system of the form A*X = B or A**T*X = B,
                 where A is a triangular band matrix of order N, and B  is  an
                 N-by-NRHS matrix

       dtfsm  - solve a matrix equation (one operand is a triangular matrix in
                 RFP format)

       dtftri - compute the inverse of a triangular matrix  A  stored  in  RFP
                 format

       dtfttp - copy a triangular matrix from the rectangular full packed for-
                 mat (TF) to the standard packed format (TP)

       dtfttr - copy a triangular matrix from the rectangular full packed for-
                 mat (TF) to the standard full format (TR)

       dtgevc  -  compute  the right and/or left generalized eigenvectors of a
                 pair  of  real  upper  triangular  matrices  (A,B)  that  was
                 obtained from the generalized Schur factorization of an orig-
                 inal pair of real nonsymmetric matrices (AO,BO)

       dtgexc - reorder the generalized Schur decomposition of a  real  matrix
                 pair  using  an orthogonal or unitary equivalence transforma-
                 tion

       dtgsen - reorder the generalized real Schur  decomposition  of  a  real
                 matrix pair (A, B), so that a selected cluster of eigenvalues
                 appears in the leading diagonal blocks of  the  upper  quasi-
                 triangular matrix A and the upper triangular B

       dtgsja - compute the generalized singular value decomposition (GSVD) of
                 two real upper triangular (or trapezoidal) matrices A and B

       dtgsna - estimate reciprocal condition numbers for specified  eigenval-
                 ues  and/or  eigenvectors of a matrix pair (A, B) in general-
                 ized real  Schur  canonical  form  (or  of  any  matrix  pair
                 (Q*A*Z',  Q*B*Z')  with orthogonal matrices Q and Z, where Z'
                 denotes the transpose of Z

       dtgsyl - solve the generalized Sylvester equation

       dtpcon - estimate the reciprocal of the condition number  of  a  packed
                 triangular  matrix  A,  in either the 1-norm or the infinity-
                 norm

       dtpmqrt - apply a real orthogonal matrix Q obtained from a "triangular-
                 pentagonal"  real  block reflector H to a general real matrix
                 C, which consists of two blocks

       dtpmv - perform one of the matrix-vector operations x := A*x, or  x  :=
                 A'*x

       dtpqrt  - compute a blocked QR factorization of a real "triangular-pen-
                 tagonal" matrix C, which is composed of a triangular block  A
                 and  pentagonal  block B, using the compact WY representation
                 for Q

       dtpqrt2 - compute a QR factorization of a real or complex  "triangular-
                 pentagonal"  matrix,  which is composed of a triangular block
                 and a pentagonal block, using the compact  WY  representation
                 for Q

       dtprfs  -  provide  error  bounds  and backward error estimates for the
                 solution to a system of linear equations  with  a  triangular
                 packed coefficient matrix

       dtpsv - solve one of the systems of equations A*x = b, or A'*x = b

       dtptri - compute the inverse of a real upper or lower triangular matrix
                 A stored in packed format

       dtptrs -  solve a triangular system of the form A*X = B or A**T*X =  B,
                 where  A  is  a triangular matrix of order N stored in packed
                 format, and B is an N-by-NRHS matrix

       dtpttf - copy a triangular matrix from the standard packed format  (TP)
                 to the rectangular full packed format (TF)

       dtpttr  - copy a triangular matrix from the standard packed format (TP)
                 to the standard full format (TR)

       dtrans - transpose and scale source matrix

       dtrcon - estimate the reciprocal of the condition number of a  triangu-
                 lar matrix A, in either the 1-norm or the infinity-norm

       dtrevc - compute some or all of the right and/or left eigenvectors of a
                 real upper quasi-triangular matrix T

       dtrexc - reorder the real Schur factorization of  a  real  matrix  A  =
                 Q*T*Q**T, so that the diagonal block of T with row index IFST
                 is moved to row ILST

       dtrmm - perform one of the matrix-matrix  operationsB  :=  alpha*op(  A
                 )*B, or B := alpha*B*op( A )

       dtrmv  -  perform one of the matrix-vector operations x := A*x, or x :=
                 A'*x

       dtrrfs - provide error bounds and  backward  error  estimates  for  the
                 solution  to  a  system of linear equations with a triangular
                 coefficient matrix

       dtrsen - reorder the real Schur factorization of  a  real  matrix  A  =
                 Q*T*Q**T,  so  that a selected cluster of eigenvalues appears
                 in the leading diagonal blocks of the upper  quasi-triangular
                 matrix T,

       dtrsm  -  solve  one  of the matrix equations   op( A )*X = alpha*B, or
                 X*op( A ) = alpha*B

       dtrsna - estimate reciprocal condition numbers for specified  eigenval-
                 ues  and/or right eigenvectors of a real upper quasi-triangu-
                 lar matrix T (or of any matrix Q*T*Q**T with Q orthogonal)

       dtrsv - solve one of the systems of equations A*x = b, or A'*x = b

       dtrsyl - solve the real Sylvester matrix equation

       dtrti2 - compute the inverse of a real upper or lower triangular matrix

       dtrtri - compute the inverse of a real upper or lower triangular matrix
                 A

       dtrtrs - solve a triangular system of the form A*X = B or A**T*X  =  B,
                 where  A is a triangular matrix of order N, and B is an N-by-
                 NRHS matrix

       dtrttf - copy a triangular matrix from the standard full format (TR) to
                 the rectangular full packed format (TF)

       dtrttp - copy a triangular matrix from the standard full format (TR) to
                 the standard packed format (TP)

       dtzrqf - routine is deprecated and has been replaced by routine DTZRZF

       dtzrzf - reduce the M-by-N ( M<=N ) real upper trapezoidal matrix A  to
                 upper triangular form by means of orthogonal transformations

       dvbrmm - variable block sparse row format matrix-matrix multiply

       dvbrsm - variable block sparse row format triangular solve

       dwiener - perform Wiener deconvolution of two signals

       dzasum - Return the sum of the absolute values of a vector x.

       dznrm2 - Return the Euclidian norm of a vector.

       ezfftb  -  computes  a periodic sequence from its Fourier coefficients.
                 EZFFTB is a simplified but slower version of RFFTB.

       ezfftf - computes the Fourier  coefficients  of  a  periodic  sequence.
                 EZFFTF is a simplified but slower version of RFFTF.

       ezffti  - initializes the array WSAVE, which is used in both EZFFTF and
                 EZFFTB.

       fft - Fast Fourier transform subroutines

       gen_custom - extract necessary routines from an archive library to cre-
                 ate a purpose-built library.

       getmsg - Open a catalog file and display the requested message

       gscon  -  estimates the reciprocal of the condition number of a general
                 real matrix A, in either the  1-norm  or  the  infinity-norm,
                 using  the  LU  factorization  computed  by  SuperLU  routine
                 sgstrf.

       gsequ - computes row and column scalings intended to equilibrate an  M-
                 by-N sparse matrix A and reduce its condition number.

       gsrfs  - improves the computed solution to a system of linear equations
                 and provides error bounds and backward  error  estimates  for
                 the solution.  It is a SuperLU routine.

       gssv  -  solves a system of linear equations A*X=B using the LU factor-
                 ization from sgstrf.

       gssvx - solves the system of linear equations A*X=B  or  A'*X=B,  using
                 the LU factorization from sgstrf(). Error bounds on the solu-
                 tion and a condition estimate are also provided.

       gstrf - computes an LU factorization of a general sparse m-by-n  matrix
                 A using partial pivoting with row interchanges.

       gstrs  -  solves  a  system  of linear equations A*X=B or A'*X=B with A
                 sparse and B dense, using the LU  factorization  computed  by
                 sgstrf.

       icamax - return the index of the element with largest absolute value.

       idamax - return the index of the element with largest absolute value.

       iladiag - translated from a character string specifying if a matrix has
                 unit diagonal or not to the relevant BLAST-specified  integer
                 constant

       ilaenv  -  The  name of the calling subroutine, in either upper case or
                 lower case.

       ilaprec - translate from a character string specifying an  intermediate
                 precision to the relevant BLAST-specified integer constant

       ilatrans - translate from a character string specifying a transposition
                 operation to the relevant BLAST-specified integer constant

       ilauplo - translate from a character  string  specifying  a  upper-  or
                 lower-triangular matrix to the relevant BLAST-specified inte-
                 ger constant

       ilaver - return the Lapack version Arguments

       infinite_norm_error - A utility function of  the  SuperLU  solver  that
                 computes  the  infinity-norm  of an array of vectors that are
                 approximations to the exact solution vector.

       intro - Introduction to Oracle Solaris Studio Performance Library func-
                 tions and subroutines

       isamax - return the index of the element with largest absolute value.

       izamax - return the index of the element with largest absolute value.

       langs  -  returns  the value of the one-norm, or the Frobenius-norm, or
                 the infinity-norm, or the element with largest absolute value
                 of a general real matrix A in sparse format.

       laqgs  -  a  SuperLU function that equilibrates a general sparse M by N
                 matrix A.

       libsunperf - Introduction to Oracle Solaris Studio Performance  Library
                 functions and subroutines

       lsame  -  returns  .TRUE.  if CA is the same letter as CB regardless of
                 case

       rfft2b - compute a periodic sequence  from  its  Fourier  coefficients.
                 The  RFFT  operations  are  unnormalized, so a call of RFFT2F
                 followed by a call of RFFT2B will multiply the input sequence
                 by M*N.

       rfft2f  - compute the Fourier coefficients of a periodic sequence.  The
                 RFFT operations are unnormalized, so a call  of  RFFT2F  fol-
                 lowed by a call of RFFT2B will multiply the input sequence by
                 M*N.

       rfft2i - initialize the array WSAVE, which is used in both the  forward
                 and backward transforms.

       rfft3b - compute a periodic sequence from its Fourier coefficients. The
                 RFFT operations are unnormalized, so a call  of  RFFT3F  fol-
                 lowed by a call of RFFT3B will multiply the input sequence by
                 M*N*K.

       rfft3f - compute the Fourier coefficients of a real periodic  sequence.
                 The  RFFT  operations  are  unnormalized, so a call of RFFT3F
                 followed by a call of RFFT3B will multiply the input sequence
                 by M*N*K.

       rfft3i  -  initialize the array WSAVE, which is used in both RFFT3F and
                 RFFT3B.

       rfftb - compute a periodic sequence from its Fourier coefficients.  The
                 RFFT operations are unnormalized, so a call of RFFTF followed
                 by a call of RFFTB will multiply the input sequence by N.

       rfftf - compute the Fourier coefficients of a periodic  sequence.   The
                 FFT  operations are unnormalized, so a call of RFFTF followed
                 by a call of RFFTB will multiply the input sequence by N.

       rffti - initialize the array WSAVE, which is used  in  both  RFFTF  and
                 RFFTB.

       rfftopt - compute the length of the closest fast FFT

       sCopy_CompCol_Matrix  -  A  utility  C  function  in the serial SuperLU
                 solver that copies one SuperMatrix into another.

       sCreate_CompCol_Matrix - A utility C function  in  the  serial  SuperLU
                 solver that creates a SuperMatrix in compressed sparse column
                 format (also known as the Harwell-Boeing format).

       sCreate_CompRow_Matrix - A utility C function  in  the  serial  SuperLU
                 solver  that  creates  a SuperMatrix in compressed sparse row
                 format.

       sCreate_Dense_Matrix - A utility  C  function  in  the  serial  SuperLU
                 solver that creates a SuperMatrix in dense format.

       sCreate_SuperNode_Matrix  -  A utility C function in the serial SuperLU
                 solver that creates a SuperMatrix in supernodal format.

       sPrintPerf - A utility function of the SuperLU solver that prints  sta-
                 tistics collected by the computational routines.

       sQuerySpace  - A inquiry function that provides information on the mem-
                 ory statistics of the SuperLU solver.

       sasum - Return the sum of the absolute values of a vector x.

       saxpy - compute y := alpha * x + y

       saxpyi - Compute y := alpha * x + y

       sbbcsd - compute the CS decomposition of an orthogonal matrix in  bidi-
                 agonal-block form

       sbcomm - block coordinate matrix-matrix multiply

       sbdimm - block diagonal format matrix-matrix multiply

       sbdism -  block diagonal format triangular solve

       sbdsdc - compute the singular value decomposition (SVD) of a real N-by-
                 N (upper or lower) bidiagonal matrix B

       sbdsqr - compute the singular value decomposition (SVD) of a real N-by-
                 N (upper or lower) bidiagonal matrix B

       sbelmm - block Ellpack format matrix-matrix multiply

       sbelsm - block Ellpack format triangular solve

       sbscmm - block sparse column matrix-matrix multiply

       sbscsm - block sparse column format triangular solve

       sbsrmm - block sparse row format matrix-matrix multiply

       sbsrsm - block sparse row format triangular solve

       scasum - Return the sum of the absolute values of a vector x.

       scnrm2 - Return the Euclidian norm of a vector.

       scnvcor - compute the convolution or correlation of real vectors

       scnvcor2 - compute the convolution or correlation of real matrices

       scoomm - coordinate matrix-matrix multiply

       scopy - copy x to y

       scscmm - compressed sparse column format matrix-matrix multiply

       scscsm - compressed sparse column format triangular solve

       scsrmm - compressed sparse row format matrix-matrix multiply

       scsrsm - compressed sparse row format triangular solve

       sdiamm - diagonal format matrix-matrix multiply

       sdiasm - diagonal format triangular solve

       sdisna  - compute the reciprocal condition numbers for the eigenvectors
                 of a real symmetric or complex Hermitian matrix  or  for  the
                 left or right singular vectors of a general m-by-n matrix

       sdot - compute the dot product of two vectors x and y.

       sdoti - Compute the indexed dot product.

       sdsdot  -  compute  a constant plus the double precision dot product of
                 two single precision vectors x and y

       second - return the user time for a process in seconds

       sellmm - Ellpack format matrix-matrix multiply

       sellsm - Ellpack format triangular solve

       set_default_options - C function that sets to  default  parameters  the
                 options  that  control  the  behavior  of  the serial SuperLU
                 solver.

       sfftc - initialize the trigonometric weight and factor tables  or  com-
                 pute the forward Fast Fourier Transform of a real sequence.

       sfftc2  - initialize the trigonometric weight and factor tables or com-
                 pute the two-dimensional forward Fast Fourier Transform of  a
                 two-dimensional real array.

       sfftc3  - initialize the trigonometric weight and factor tables or com-
                 pute the three-dimensional forward Fast Fourier Transform  of
                 a three-dimensional complex array.

       sfftcm  - initialize the trigonometric weight and factor tables or com-
                 pute the one-dimensional forward Fast Fourier Transform of  a
                 set of real data sequences stored in a two-dimensional array.

       sgbbrd - reduce a real general m-by-n band matrix A to upper bidiagonal
                 form B by an orthogonal transformation

       sgbcon - estimate the reciprocal of the condition number of a real gen-
                 eral band matrix A, in either the  1-norm  or  the  infinity-
                 norm, using the LU factorization computed by SGBTRF

       sgbequ  - compute row and column scalings intended to equilibrate an M-
                 by-N band matrix A and reduce its condition number

       sgbequb - compute row and column scalings intended to equilibrate an M-
                 by-N matrix A and reduce its condition number

       sgbmv  -  perform  one  of the matrix-vector operationsy := alpha*A*x +
                 beta*y or y := alpha*A'*x + beta*y

       sgbrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix  is  banded, and provide error
                 bounds and backward error estimates for the solution

       sgbrfsx - improve the computed solution to a system of linear equations
                 and provide error bounds and backward error estimates for the
                 solution

       sgbsv - compute the solution to  a  real  system  of  linear  equations
                 A*X=B,  where A is a band matrix of order N with KL subdiago-
                 nals and KU superdiagonals, and X and B are N-by-NRHS  matri-
                 ces

       sgbsvx  -  use  the  LU factorization to compute the solution to a real
                 system of linear  equations  A*X=B,  A**T*X=B,  or  A**H*X=B,
                 where A is a band matrix

       sgbsvxx  - compute the solution to system of linear equations A * X = B
                 for genaral band matrices

       sgbtf2 - compute an LU factorization of a real  m-by-n  band  matrix  A
                 using partial pivoting with row interchanges

       sgbtrf  -  compute  an  LU factorization of a real m-by-n band matrix A
                 using partial pivoting with row interchanges

       sgbtrs - solve a system of linear equations A*X=B or A'*X=B with a gen-
                 eral  band  matrix  A  using the LU factorization computed by
                 SGBTRF

       sgebak - form the right or left eigenvectors of a real  general  matrix
                 by  backward  transformation  on the computed eigenvectors of
                 the balanced matrix output by SGEBAL

       sgebal - balance a general real matrix A

       sgebd2 - reduce a general matrix to bidiagonal form using an  unblocked
                 algorithm

       sgebrd  - reduce a general real M-by-N matrix A to upper or lower bidi-
                 agonal form B by an orthogonal transformation

       sgecon - estimate the reciprocal of the condition number of  a  general
                 real  matrix  A,  in  either the 1-norm or the infinity-norm,
                 using the LU factorization computed by SGETRF

       sgeequ - compute row and column scalings intended to equilibrate an  M-
                 by-N matrix A and reduce its condition number

       sgeequb - compute row and column scalings intended to equilibrate an M-
                 by-N matrix A and reduce its condition number

       sgees - compute for an N-by-N real nonsymmetric matrix A, the eigenval-
                 ues,  the  real  Schur form T, and, optionally, the matrix of
                 Schur vectors Z

       sgeesx - compute for an N-by-N real nonsymmetric matrix A,  the  eigen-
                 values, the real Schur form T, and, optionally, the matrix of
                 Schur vectors Z

       sgeev - compute for an N-by-N real nonsymmetric matrix A, the eigenval-
                 ues and, optionally, the left and/or right eigenvectors

       sgeevx  -  compute for an N-by-N real nonsymmetric matrix A, the eigen-
                 values and, optionally, the left and/or right eigenvectors

       sgegs - routine is deprecated and has been replaced by routine SGGES

       sgegv - routine is deprecated and has been replaced by routine SGGEV

       sgehd2 - reduce a general square matrix to upper Hessenberg form  using
                 an unblocked algorithm

       sgehrd  -  reduce a real general matrix A to upper Hessenberg form H by
                 an orthogonal similarity transformation

       sgejsv - compute the singular value decomposition (SVD) of a real M-by-
                 N matrix A, where M >= N

       sgelq2  -  compute the LQ factorization of a general rectangular matrix
                 using an unblocked algorithm

       sgelqf - compute an LQ factorization of a real M-by-N matrix A

       sgels - solve overdetermined or  underdetermined  real  linear  systems
                 involving an M-by-N matrix A, or its transpose, using a QR or
                 LQ factorization of A

       sgelsd - compute the minimum-norm  solution  to  a  real  linear  least
                 squares problem

       sgelss  -  compute  the  minimum  norm  solution to a real linear least
                 squares problem

       sgelsx - routine is deprecated and has been replaced by routine SGELSY

       sgelsy - compute the minimum-norm  solution  to  a  real  linear  least
                 squares problem

       sgemm  -  perform  one  of the matrix-matrix operationsC := alpha*op( A
                 )*op( B ) + beta*C

       sgemqrt - overwrites the general real M-by-N matrix C with Q  C,  C  Q,
                 Q**T C, or C Q**T depe nding on values of SIDE and TRANS

       sgemv  -  perform  one  of the matrix-vector operationsy := alpha*A*x +
                 beta*y or y := alpha*A'*x + beta*y

       sgeql2 - compute the QL factorization of a general  rectangular  matrix
                 using an unblocked algorithm

       sgeqlf - compute a QL factorization of a real M-by-N matrix A

       sgeqp3 - compute a QR factorization with column pivoting of a matrix A

       sgeqpf - routine is deprecated and has been replaced by routine SGEQP3

       sgeqr2  - computes the QR factorization of a general rectangular matrix
                 using an unblocked algorithm.

       sgeqr2p - computes the QR factorization of a general rectangular matrix
                 with  non-negative diagonal elements using an unblocked algo-
                 rithm.

       sgeqrf - compute a QR factorization of a real M-by-N matrix A

       sgeqrfp - compute a QR factorization of a real M-by-N matrix A: A = Q *
                 R

       sgeqrt  -  compute a blocked QR factorization of a real M-by-N matrix A
                 using the compact WY representation of Q

       sgeqrt2 - compute a QR factorization of a general real matrix using the
                 compact WY representation of Q

       sgeqrt3  -  recursively  compute  a  QR factorization of a general real
                 matrix using the compact WY representation of Q

       sger - perform the rank 1 operation A := alpha*x*y' + A

       sgerfs - improve the computed solution to a system of linear  equations
                 and provide error bounds and backward error estimates for the
                 solution

       sgerfsx - improve the computed solution to a system of linear equations
                 and provide error bounds and backward error estimates for the
                 solution

       sgerq2 - computes the RQ factorization of a general rectangular  matrix
                 using an unblocked algorithm

       sgerqf - compute an RQ factorization of a real M-by-N matrix A

       sgesdd - compute the singular value decomposition (SVD) of a real M-by-
                 N matrix A, optionally computing the left and right  singular
                 vectors

       sgesv  -  compute  the  solution  to  a real system of linear equations
                 A*X=B, where A is an N-by-N matrix and X and B are  N-by-NRHS
                 matrices

       sgesvd - compute the singular value decomposition (SVD) of a real M-by-
                 N matrix A, optionally computing the left and/or right singu-
                 lar vectors

       sgesvj - compute the singular value decomposition (SVD) of a real M-by-
                 N matrix A, where M >= N

       sgesvx - use the LU factorization to compute the  solution  to  a  real
                 system  of linear equations  A*X=B, where A is an N-by-N gen-
                 eral matrix

       sgesvxx - compute the solution to system of linear equations A*X=B  for
                 general matrices

       sgetf2 - compute an LU factorization of a general m-by-n matrix A using
                 partial pivoting with row interchanges

       sgetrf - compute an LU factorization of a general M-by-N matrix A using
                 partial pivoting with row interchanges

       sgetri  -  compute  the  inverse of a matrix using the LU factorization
                 computed by SGETRF

       sgetrs - solve a system of linear equations  A * X = B or A' *  X  =  B
                 with  a  general  N-by-N  matrix A using the LU factorization
                 computed by SGETRF

       sggbak - form the right or left eigenvectors of a real generalized  ei-
                 genvalue problem A*x = lambda*B*x, by backward transformation
                 on the computed eigenvectors of the balanced pair of matrices
                 output by SGGBAL

       sggbal - balance a pair of general real matrices (A,B)

       sgges - compute for a pair of N-by-N real nonsymmetric matrices (A,B),

       sggesx - compute for a pair of N-by-N real nonsymmetric matrices (A,B),
                 the generalized eigenvalues, the real Schur form (S,T),  and,
                 and, optionally, the left and/or right matrices of Schur vec-
                 tors

       sggev - compute for a pair of N-by-N real nonsymmetric matrices (A,B)

       sggevx - compute for a pair of N-by-N real nonsymmetric matrices  (A,B)
                 the generalized eigenvalues, and, optionally, the left and/or
                 right generalized eigenvectors

       sggglm - solve a general Gauss-Markov linear model (GLM) problem

       sgghrd - reduce a pair of real matrices (A,B) to generalized upper Hes-
                 senberg  form  using orthogonal transformations, where A is a
                 general matrix and B is upper triangular

       sgglse - solve the  linear  equality-constrained  least  squares  (LSE)
                 problem

       sggqrf  -  compute a generalized QR factorization of an N-by-M matrix A
                 and an N-by-P matrix B.

       sggrqf - compute a generalized RQ factorization of an M-by-N  matrix  A
                 and a P-by-N matrix B

       sggsvd - compute the generalized singular value decomposition (GSVD) of
                 an M-by-N real matrix A and P-by-N real matrix B

       sggsvp - compute orthogonal matrices

       sgscon - estimates the reciprocal of the condition number of a  general
                 real  matrix  A,  in  either the 1-norm or the infinity-norm,
                 using  the  LU  factorization  computed  by  SuperLU  routine
                 sgstrf.

       sgsequ - computes row and column scalings intended to equilibrate an M-
                 by-N sparse matrix A and reduce its condition number.

       sgsrfs - improves the computed solution to a system of linear equations
                 and  provides  error  bounds and backward error estimates for
                 the solution.  It is a SuperLU routine.

       sgssco - General sparse solver condition number estimate.

       sgssda - Deallocate working storage for the general sparse solver.

       sgssfa - General sparse solver numeric factorization.

       sgssfs - General sparse solver one call interface.

       sgssin - Initialize the general sparse solver.

       sgssor - General sparse solver ordering and symbolic factorization.

       sgssps - Print general sparse solver statics.

       sgssrp - Return permutation used by the general sparse solver.

       sgsssl - Solve routine for the general sparse solver.

       sgssuo - Provide general sparse solvers SPSOLVE and  SuperLU   a  user-
                 supplied permutation for ordering.

       sgssv  - solves a system of linear equations A*X=B using the LU factor-
                 ization from sgstrf.

       sgssvx - solves the system of linear equations A*X=B or  A'*X=B,  using
                 the LU factorization from sgstrf(). Error bounds on the solu-
                 tion and a condition estimate are also provided.

       sgstrf - computes an LU factorization of a general sparse m-by-n matrix
                 A using partial pivoting with row interchanges.

       sgstrs  -  solves  a  system of linear equations A*X=B or A'*X=B with A
                 sparse and B dense, using the LU  factorization  computed  by
                 sgstrf.

       sgsvj0 - pre-processor for the routine sgesvj

       sgsvj1  -  pre-processor for the routine sgesvj, apply Jacobi rotations
                 targeting only particular pivots

       sgtcon - estimate the reciprocal of the  condition  number  of  a  real
                 tridiagonal  matrix  A using the LU factorization as computed
                 by SGTTRF

       sgthr - Gathers specified elements from y into x.

       sgthrz - Gather and zero.

       sgtrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix  is tridiagonal, provide error
                 bounds and backward error estimates for the solution

       sgtsv - solve the equation A*X=B, where  A  is  an  N-by-N  tridiagonal
                 matrix, by Gaussian elimination with partial pivoting

       sgtsvx  -  use  the  LU factorization to compute the solution to a real
                 system of linear equations A*X = B or A**T*X = B, where A  is
                 a  tridiagonal  matrix  of  order N and X and B are N-by-NRHS
                 matrices

       sgttrf - compute an LU factorization of a  real  tridiagonal  matrix  A
                 using elimination with partial pivoting and row interchanges

       sgttrs  - solve one of the systems of equations A*X=B or A'*X=B, with a
                 tridiagonal matrix A using the LU factorization  computed  by
                 SGTTRF

       sgtts2  -  solve a system of linear equations with a tridiagonal matrix
                 using the LU factorization computed by sgttrf

       shgeqz - implement a single-/double-shift version of the QZ method  for
                 finding   the   generalized  eigenvalues   w(j)=(ALPHAR(j)  +
                 i*ALPHAI(j))/BETAR(j) of the equation   det( A-w(i) B )  =  0
                 In addition, the pair A,B may be reduced to generalized Schur
                 form

       shsein - use inverse iteration to  find  specified  right  and/or  left
                 eigenvectors of a real upper Hessenberg matrix H

       shseqr  -  compute  the eigenvalues of a real upper Hessenberg matrix H
                 and, optionally, the matrices T and Z from the Schur decompo-
                 sition  H  =  Z  T Z**T, where T is an upper quasi-triangular
                 matrix (the Schur form), and Z is the  orthogonal  matrix  of
                 Schur vectors

       sinfinite_norm_error  -  A  utility function of the SuperLU solver that
                 computes the infinity-norm of an array of  vectors  that  are
                 approximations to the exact solution vector.

       sinqb  - synthesize a Fourier sequence from its representation in terms
                 of a sine series with odd wave numbers.  The SINQ  operations
                 are  unnormalized  inverses of themselves, so a call to SINQF
                 followed by a call to SINQB will multiply the input  sequence
                 by 4 * N.

       sinqf  -  compute the Fourier coefficients in a sine series representa-
                 tion with only  odd  wave  numbers.The  SINQ  operations  are
                 unnormalized  inverses of themselves, so a call to SINQF fol-
                 lowed by a call to SINQB will multiply the input sequence  by
                 4 * N.

       sinqi  -  initialize  the array xWSAVE, which is used in both SINQF and
                 SINQB.

       sint - compute the discrete Fourier sine transform of an odd  sequence.
                 The  SINT transforms are unnormalized inverses of themselves,
                 so a call of SINT followed by another call of SINT will  mul-
                 tiply the input sequence by 2 * (N+1).

       sinti - initialize the array WSAVE, which is used in subroutine SINT.

       sjadmm - Jagged diagonal matrix-matrix multiply (modified Ellpack)

       sjadrp - right permutation of a jagged diagonal matrix

       sjadsm - Jagged-diagonal format triangular solve

       sla_gbamv - perform a matrix-vector operation to calculate error bounds

       sla_gbrcond - estimate the Skeel condition number for a general  banded
                 matrix

       sla_gbrfsx_extended - improve the computed solution to a system of lin-
                 ear equations  for  general  banded  matrices  by  performing
                 extra-precise  iterative  refinement and provide error bounds
                 and backward error estimates for the solution

       sla_gbrpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U) for a general banded matrix

       sla_geamv  -  compute a matrix-vector product using a general matrix to
                 calculate error bounds

       sla_gercond - estimate the Skeel condition number for a general matrix

       sla_gerfsx_extended - improve the computed solution to a system of lin-
                 ear  equations  for general matrices by performing extra-pre-
                 cise iterative refinement and provide error bounds and  back-
                 ward error estimates for the solution

       sla_gerpvgrw  -  compute  the  reciprocal pivot growth factor using the
                 "max absolute element" norm

       sla_lin_berr - compute a component-wise relative backward error

       sla_porcond - estimate the Skeel condition number for a symmetric posi-
                 tive-definite matrix

       sla_porfsx_extended - improve the computed solution to a system of lin-
                 ear equations for symmetric  or  Hermitian  positive-definite
                 matrices by performing extra-precise iterative refinement and
                 provide error bounds and backward  error  estimates  for  the
                 solution

       sla_porpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U) for a symmetric or  Hermitian  positive-defi-
                 nite matrix

       sla_syamv  -  compute a matrix-vector product using a symmetric indefi-
                 nite matrix to calculate error bounds

       sla_syrcond - estimate the  Skeel  condition  number  for  a  symmetric
                 indefinite matrix

       sla_syrfsx_extended - improve the computed solution to a system of lin-
                 ear equations for symmetric indefinite matrices by performing
                 extra-precise  iterative  refinement and provide error bounds
                 and backward error estimates for the solution

       sla_syrpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U) for a symmetric indefinite matrix

       sla_wwaddw - add a vector into a doubled-single vector

       slaed0  -  compute all eigenvalues and corresponding eigenvectors of an
                 unreduced symmetric tridiagonal matrix using the  divide  and
                 conquer method. Used by sstedc

       slaed1  -  compute  the  updated eigensystem of a diagonal matrix after
                 modification by a rank-one symmetric matrix. Used by  sstedc,
                 when the original matrix is tridiagonal

       slaed2  - merge eigenvalues and deflates secular equation; used by sst-
                 edc when the original matrix is tridiagonal

       slaed3 - find the roots of the secular equation and updates the  eigen-
                 vectors; used by sstedc when the original matrix is tridiago-
                 nal

       slaed4 - is used by sstedc. Finds a single root of the secular equation

       slaed5 - is used by sstedc. Solves the 2-by-2 secular equation

       slaed6  - is used by sstedc. Compute one Newton step in solution of the
                 secular equation

       slaed7 - compute the updated eigensystem of  a  diagonal  matrix  after
                 modification  by a rank-one symmetric matrix. Used by sstedc,
                 when the original matrix is dense

       slaed8 - merge eigenvalues and deflates secular equation. Used by  sst-
                 edc, when the original matrix is dense

       slaed9  - is used by sstedc. Find the roots of the secular equation and
                 updates the eigenvectors. Used when the  original  matrix  is
                 dense

       slaeda  - is used by sstedc. Compute the Z vector determining the rank-
                 one modification of the diagonal matrix. Used when the origi-
                 nal matrix is dense

       slagtf - factorize the matrix (T-lambda*I), where T is an n by n tridi-
                 agonal matrix and lambda is a scalar, as T-lambda*I = PLU

       slals0 - apply back multiplying factors in solving  the  least  squares
                 problem using divide and conquer SVD approach. Used by sgelsd

       slalsa - compute the SVD of the coefficient  matrix  in  compact  form.
                 Used by sgelsd

       slalsd  -  use the singular value decomposition of A to solve the least
                 squares problem

       slamch - Determines single precision machine parameters.

       slamrg - will create a permutation list which will merge  the  elements
                 of  A  (which  is  composed of two independently sorted sets)
                 into a single set which is sorted in ascending order

       slangs - returns the value of the one-norm, or the  Frobenius-norm,  or
                 the infinity-norm, or the element with largest absolute value
                 of a general real matrix A in sparse format.

       slansf - return the value of the one norm, or the  Frobenius  norm,  or
                 the  infinity  norm, or the element of largest absolute value
                 of a real symmetric matrix A in RFP format

       slaqgs - a SuperLU function that equilibrates a general sparse M  by  N
                 matrix A.

       slarscl2 - perform reciprocal diagonal scaling on a vector

       slartgs  -  generate  a plane rotation designed to introduce a bulge in
                 implicit QR iteration for the bidiagonal SVD problem

       slarz - apply a real elementary reflector H to a real M-by-N matrix  C,
                 from either the left or the right

       slarzb - apply a real block reflector H or its transpose H**T to a real
                 distributed M-by-N C from the left or the right

       slarzt - form the triangular factor T of a real block  reflector  H  of
                 order  >  n,  which  is  defined as a product of k elementary
                 reflectors

       slascl2 - perform diagonal scaling on a vector

       slasq1 - compute the  singular  values  of  a  real  square  bidiagonal
                 matrix. Used by sbdsqr

       slasq2 - compute all the eigenvalues of a real symmetric positive defi-
                 nite tridiagonal matrix (high relative accuracy)

       slasq3 - check for deflation, compute a shift and calls dqds.  Used  by
                 sbdsqr

       slasq4 - compute an approximation to the smallest eigenvalue using val-
                 ues of d from the previous transform. Used by sbdsqr

       slasq5 - compute one dqds transform in ping-pong form. Used  by  sbdsqr
                 and sstegr

       slasq6  -  compute  one dqd transform in ping-pong form. Used by sbdsqr
                 and sstegr

       slasrt - the numbers in D in increasing order  (if  ID  =  'I')  or  in
                 decreasing order (if ID = 'D' )

       slasyf  -  compute  a partial factorization of a real symmetric matrix,
                 using the Bunch-Kaufman diagonal pivoting method

       slasyf_rook - compute a  partial  factorization  of  a  real  symmetric
                 matrix using the bounded Bunch-Kaufman ("rook") diagonal piv-
                 oting method

       slatrz - factor an upper trapezoidal  matrix  by  means  of  orthogonal
                 transformations

       slatzm - routine is deprecated and has been replaced by routine SORMRZ

       snrm2 - Return the Euclidian norm of a vector.

       sopgtr  -  generate  a real orthogonal matrix Q which is defined as the
                 product of n-1 elementary reflectors  H(i)  of  order  n,  as
                 returned by SSPTRD using packed storage

       sopmtr  -  overwrite the general real M-by-N matrix C with   SIDE = 'L'
                 SIDE = 'R' TRANS = 'N'

       sorbdb - simultaneously bidiagonalize the blocks of  an  M-by-M  parti-
                 tioned orthogonal matrix

       sorbdb1  - simultaneously bidiagonalize the blocks of a tall and skinny
                 matrix with orthonomal columns

       sorbdb2 - simultaneously bidiagonalize the blocks of a tall and  skinny
                 matrix with orthonomal columns

       sorbdb3  - simultaneously bidiagonalize the blocks of a tall and skinny
                 matrix with orthonomal columns

       sorbdb4 - simultaneously bidiagonalize the blocks of a tall and  skinny
                 matrix with orthonomal columns

       sorbdb5 - orthogonalize one column vector with respect to the orthonor-
                 mal columns of another column vector

       sorbdb6 - orthogonalize one column vector with respect to the orthonor-
                 mal columns of another column vector

       sorcsd - compute the CS decomposition of an M-by-M partitioned orthogo-
                 nal matrix

       sorcsd2by1 - compute the CS decomposition  of  an  M-by-Q  matrix  with
                 orthonormal  columns  that has been partitioned into a 2-by-1
                 block structure

       sorg2l - generate an m by n real matrix Q with orthonormal columns,

       sorg2r - generate an m by n real matrix Q with orthonormal columns,

       sorgbr - generate one of the real orthogonal matrices Q or P**T  deter-
                 mined  by  SGEBRD when reducing a real matrix A to bidiagonal
                 form

       sorghr - generate a real orthogonal matrix Q which is  defined  as  the
                 product  of  IHI-ILO  elementary  reflectors  of  order N, as
                 returned by SGEHRD

       sorgl2 - generate an m by n real matrix Q with orthonormal rows,

       sorglq - generate an M-by-N real matrix Q with orthonormal rows,

       sorgql - generate an M-by-N real matrix Q with orthonormal columns,

       sorgqr - generate an M-by-N real matrix Q with orthonormal columns,

       sorgr2 - generate an m by n real matrix Q with orthonormal rows,

       sorgrq - generate an M-by-N real matrix Q with orthonormal rows,

       sorgtr - generate a real orthogonal matrix Q which is  defined  as  the
                 product  of n-1 elementary reflectors of order N, as returned
                 by SSYTRD

       sorm2l - multiply a general matrix by the orthogonal matrix from  a  QL
                 factorization determined by sgeqlf (unblocked algorithm)

       sorm2r  -  multiply a general matrix by the orthogonal matrix from a QR
                 factorization determined by sgeqrf (unblocked algorithm)

       sormbr - overwrites the general real M-by-N matrix C with Q*C or Q**T*C
                 or C*Q**T or C*Q or P*C or P**T*C or C*P or C*P**T.

       sormhr  - overwrite the general real M-by-N matrix C with Q*C or Q**T*C
                 or C*Q**T or C*Q.

       sorml2 - multiply a general matrix by the orthogonal matrix from  a  LQ
                 factorization determined by sgelqf (unblocked algorithm)

       sormlq  - overwrite the general real M-by-N matrix C with Q*C or Q**T*C
                 or C*Q**T or C*Q.

       sormql - overwrite the general real M-by-N matrix C with Q*C or  Q**T*C
                 or C*Q**T or C*Q.

       sormqr  -  overwrite the general real M-by-N matrix C with   SIDE = 'L'
                 SIDE = 'R' TRANS = 'N'

       sormr2 - multipliy a general matrix by the orthogonal matrix from a  RQ
                 factorization determined by sgerqf (unblocked algorithm)

       sormr3  -  multiply a general matrix by the orthogonal matrix from a RZ
                 factorization determined by stzrzf (unblocked algorithm)

       sormrq - overwrite the general real M-by-N matrix C with  Q*C or Q**T*C
                 or C*Q**T or C*Q.

       sormrz  - overwrite the general real M-by-N matrix C with Q*C or Q**H*C
                 or C*Q**H or C*Q.

       sormtr - overwrite the general real M-by-N matrix C with Q*C or  Q**T*C
                 or C*Q**T or C*Q.

       sp_cgemm  -  a  SuperLU  routine that performs one of the matrix-matrix
                 operations  C := alpha*op( A )*op( B ) + beta*C  where  op(X)
                 is  one  of    op(X)  = X or op(X) = X' or op(X) = conjg(X'),
                 alpha and beta are scalars, A is  a  sparse  matrix  of  type
                 SuperMatrix,  and B and C are dense matrices, with op( A ) an
                 m by k matrix,op( B ) a  k by n matrix  and   C  an  m  by  n
                 matrix.

       sp_dgemm  -  a  SuperLU  routine that performs one of the matrix-matrix
                 operations  C := alpha*op( A )*op( B ) + beta*C  where  op(X)
                 is  one  of    op(X)  = X or op(X) = X' or op(X) = conjg(X'),
                 alpha and beta are scalars, A is  a  sparse  matrix  of  type
                 SuperMatrix,  and B and C are dense matrices, with op( A ) an
                 m by k matrix,op( B ) a  k by n matrix  and   C  an  m  by  n
                 matrix.

       sp_gemm  -  a  SuperLU  routine  that performs one of the matrix-matrix
                 operations  C := alpha*op( A )*op( B ) + beta*C  where  op(X)
                 is  one  of    op(X)  = X or op(X) = X' or op(X) = conjg(X'),
                 alpha and beta are scalars, A is  a  sparse  matrix  of  type
                 SuperMatrix,  and B and C are dense matrices, with op( A ) an
                 m by k matrix,op( B ) a  k by n matrix  and   C  an  m  by  n
                 matrix.

       sp_ienv - called by SuperLU routines to choose machine dependent param-
                 eters for the local environment. See ISPEC for a  description
                 of the parameters.

       sp_preorder - permutes the columns of the original sparse matrix.

       sp_sgemm  -  a  SuperLU  routine that performs one of the matrix-matrix
                 operations  C := alpha*op( A )*op( B ) + beta*C  where  op(X)
                 is  one  of    op(X)  = X or op(X) = X' or op(X) = conjg(X'),
                 alpha and beta are scalars, A is  a  sparse  matrix  of  type
                 SuperMatrix,  and B and C are dense matrices, with op( A ) an
                 m by k matrix,op( B ) a  k by n matrix  and   C  an  m  by  n
                 matrix.

       sp_zgemm  -  a  SuperLU  routine that performs one of the matrix-matrix
                 operations  C := alpha*op( A )*op( B ) + beta*C  where  op(X)
                 is  one  of    op(X)  = X or op(X) = X' or op(X) = conjg(X'),
                 alpha and beta are scalars, A is  a  sparse  matrix  of  type
                 SuperMatrix,  and B and C are dense matrices, with op( A ) an
                 m by k matrix,op( B ) a  k by n matrix  and   C  an  m  by  n
                 matrix.

       spbcon  -  estimate  the  reciprocal  of  the  condition number (in the
                 1-norm) of a real symmetric  positive  definite  band  matrix
                 using  the  Cholesky  factorization  A = U**T*U or A = L*L**T
                 computed by SPBTRF

       spbequ - compute row and column scalings intended to equilibrate a sym-
                 metric  positive definite band matrix A and reduce its condi-
                 tion number (with respect to the two-norm)

       spbrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix is symmetric positive definite
                 and banded, provide error bounds and backward error estimates
                 for the solution

       spbstf  -  compute  a  split Cholesky factorization of a real symmetric
                 positive definite band matrix A

       spbsv - compute the solution to  a  real  system  of  linear  equations
                 A*X=B,  where A is an N-by-N symmetric positive definite band
                 matrix and X and B are N-by-NRHS matrices

       spbsvx - use the Cholesky factorization to compute the  solution  to  a
                 real  system  of linear equations A*X=B, where A is an N-by-N
                 symmetric positive definite band matrix and X and B are N-by-
                 NRHS matrices

       spbtf2  -  compute the Cholesky factorization of a real symmetric posi-
                 tive definite band matrix A

       spbtrf - compute the Cholesky factorization of a real  symmetric  posi-
                 tive definite band matrix A

       spbtrs  -  solve  a system of linear equations A*X = B with a symmetric
                 positive definite band matrix A using the Cholesky factoriza-
                 tion A = U**T*U or A = L*L**T computed by SPBTRF

       spftrf  -  compute the Cholesky factorization of a real symmetric posi-
                 tive definite matrix A, the block version of the algorithm

       spftri - compute the inverse of  a  real  symmetric  positive  definite
                 matrix A using the Cholesky factorization computed by SPFTRF

       spftrs  -  solve  a system of linear equations A*X = B with a symmetric
                 positive definite matrix A using the  Cholesky  factorization
                 computed by SPFTRF

       spocon  -  estimate  the  reciprocal  of  the  condition number (in the
                 1-norm) of a real symmetric positive  definite  matrix  using
                 the  Cholesky factorization A = U**T*U or A = L*L**T computed
                 by SPOTRF

       spoequ - compute row and column scalings intended to equilibrate a sym-
                 metric  positive  definite  matrix A and reduce its condition
                 number (with respect to the two-norm)

       spoequb - compute row and column scalings  intended  to  equilibrate  a
                 symmetric positive definite matrix A and reduce its condition
                 number with respect to the two-norm

       sporfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient matrix is symmetric positive definite,
                 provide error bounds and backward  error  estimates  for  the
                 solution

       sporfsx - improve the computed solution to a system of linear equations
                 when the coefficient matrix is symmetric  positive  definite,
                 provide  error  bounds  and  backward error estimates for the
                 solution

       sposv - compute the solution to a real system of linear equations   A*X
                 =  B, where A is an N-by-N symmetric positive definite matrix
                 and X and B are N-by-NRHS matrices

       sposvx - use the Cholesky factorization A = U**T*U or  A  =  L*L**T  to
                 compute  the solution to a real system of linear equationsA*X
                 = B, where A is an N-by-N symmetric positive definite  matrix
                 and X and B are N-by-NRHS matrices

       sposvxx - compute the solution to a real system of linear equations A*X
                 = B, where A is an N-by-N symmetric positive definite  matrix
                 and X and B are N-by-NRHS matrices

       spotf2  -  compute the Cholesky factorization of a real symmetric posi-
                 tive definite matrix A

       spotrf - compute the Cholesky factorization of a real  symmetric  posi-
                 tive definite matrix A

       spotri  -  compute  the  inverse  of a real symmetric positive definite
                 matrix A using the Cholesky factorization A = U**T*U or  A  =
                 L*L**T computed by SPOTRF

       spotrs  -  solve  a system of linear equations A*X = B with a symmetric
                 positive definite matrix A using the Cholesky factorization A
                 = U**T*U or A = L*L**T computed by SPOTRF

       sppcon  -  estimate  the  reciprocal  of  the  condition number (in the
                 1-norm) of a real symmetric positive definite  packed  matrix
                 using  the  Cholesky  factorization  A = U**T*U or A = L*L**T
                 computed by SPPTRF

       sppequ - compute row and column scalings intended to equilibrate a sym-
                 metric  positive  definite  matrix  A  in  packed storage and
                 reduce its condition number (with respect to the two-norm)

       spprfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix is symmetric positive definite
                 and packed, provide error bounds and backward error estimates
                 for the solution

       sppsv - compute the solution to a real system of linear equations A*X =
                 B, where A is an N-by-N symmetric  positive  definite  matrix
                 stored in packed format and X and B are N-by-NRHS matrices

       sppsvx  -  use  the Cholesky factorization to compute the solution to a
                 real system of linear equations  A * X = B, where A is an  N-
                 by-N symmetric positive definite matrix stored in packed for-
                 mat and X and B are N-by-NRHS matrices

       spptrf - compute the Cholesky factorization of a real  symmetric  posi-
                 tive definite matrix A stored in packed format

       spptri  -  compute  the  inverse  of a real symmetric positive definite
                 matrix A using the Cholesky factorization A = U**T*U or  A  =
                 L*L**T computed by SPPTRF

       spptrs  -  solve  a system of linear equations A*X = B with a symmetric
                 positive definite  matrix  A  in  packed  storage  using  the
                 Cholesky  factorization  A = U**T*U or A = L*L**T computed by
                 SPPTRF

       spstf2 - compute the Cholesky factorization with complete pivoting of a
                 real symmetric positive semidefinite matrix A

       spstrf - compute the Cholesky factorization with complete pivoting of a
                 real symmetric positive semidefinite matrix A

       sptcon - compute the reciprocal of the condition number (in the 1-norm)
                 of  a  real  symmetric  positive  definite tridiagonal matrix
                 using the factorization A = L*D*L**T or A = U**T*D*U computed
                 by SPTTRF

       spteqr  -  compute  all  eigenvalues and, optionally, eigenvectors of a
                 symmetric positive definite tridiagonal matrix by first  fac-
                 toring  the  matrix  using SPTTRF, and then calling SBDSQR to
                 compute the singular values of the bidiagonal factor

       sptrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix is symmetric positive definite
                 and tridiagonal, provide  error  bounds  and  backward  error
                 estimates for the solution

       sptsv - compute the solution to a real system of linear equations A*X =
                 B, where A is an N-by-N symmetric positive definite tridiago-
                 nal matrix, and X and B are N-by-NRHS matrices

       sptsvx  - use the factorization A = L*D*L**T to compute the solution to
                 a real system of linear equations A*X = B, where A is  an  N-
                 by-N symmetric positive definite tridiagonal matrix and X and
                 B are N-by-NRHS matrices

       spttrf - compute the L*D*L' factorization of a real symmetric  positive
                 definite tridiagonal matrix A

       spttrs  -  solve  a tridiagonal system of the form  A * X = B using the
                 L*D*L' factorization of A computed by SPTTRF

       sptts2 - solve a tridiagonal system of the form  A * X =  B  using  the
                 L*D*L' factorization of A computed by SPTTRF

       srot - Apply a Given's rotation constructed by SROTG

       srotg - Construct a Given's plane rotation

       sroti -  Apply an indexed Givens rotation.

       srotm  -  Apply  a Gentleman's modified Given's rotation constructed by
                 SROTMG.

       srotmg - Construct a Gentleman's modified Given's plane rotation

       ssbev - compute all the eigenvalues and, optionally, eigenvectors of  a
                 real symmetric band matrix A

       ssbevd - compute all the eigenvalues and, optionally, eigenvectors of a
                 real symmetric band matrix A

       ssbevx - compute selected eigenvalues and, optionally, eigenvectors  of
                 a real symmetric band matrix A

       ssbgst - reduce a real symmetric-definite banded generalized eigenprob-
                 lem A*x = lambda*B*x to standard form C*y = lambda*y,

       ssbgv - compute all the eigenvalues, and optionally,  the  eigenvectors
                 of a real generalized symmetric-definite banded eigenproblem,
                 of the form A*x=(lambda)*B*x

       ssbgvd - compute all the eigenvalues, and optionally, the  eigenvectors
                 of a real generalized symmetric-definite banded eigenproblem,
                 of the form A*x=(lambda)*B*x

       ssbgvx - compute selected eigenvalues, and optionally, eigenvectors  of
                 a real generalized symmetric-definite banded eigenproblem, of
                 the form A*x=(lambda)*B*x

       ssbmv - perform the matrix-vector operationy := alpha*A*x + beta*y

       ssbtrd - reduce a real symmetric band matrix A to symmetric tridiagonal
                 form T by an orthogonal similarity transformation

       sscal - Compute y := alpha * y

       ssctr - Scatters elements from x into y

       ssfrk - perform a symmetric rank-k operation for matrix in RFP format

       sskymm - Skyline format matrix-matrix multiply

       sskysm - Skyline format triangular solve

       sspcon  -  estimate  the  reciprocal  of  the  condition number (in the
                 1-norm) of a real symmetric packed matrix A using the factor-
                 ization A = U*D*U**T or A = L*D*L**T computed by SSPTRF

       sspev  - compute all the eigenvalues and, optionally, eigenvectors of a
                 real symmetric matrix A in packed storage

       sspevd - compute all the eigenvalues and, optionally, eigenvectors of a
                 real symmetric matrix A in packed storage

       sspevx  - compute selected eigenvalues and, optionally, eigenvectors of
                 a real symmetric matrix A in packed storage

       sspgst - reduce a real symmetric-definite generalized  eigenproblem  to
                 standard form, using packed storage

       sspgv  -  compute all the eigenvalues and, optionally, the eigenvectors
                 of a real generalized symmetric-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       sspgvd  - compute all the eigenvalues, and optionally, the eigenvectors
                 of a real generalized symmetric-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       sspgvx  - compute selected eigenvalues, and optionally, eigenvectors of
                 a real generalized symmetric-definite  eigenproblem,  of  the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       sspmv - perform the matrix-vector operationy := alpha*A*x + beta*y

       sspr - perform the symmetric rank 1 operation   A := alpha*x*x' + A

       sspr2  -  perform  the  symmetric  rank 2 operation   A := alpha*x*y' +
                 alpha*y*x' + A

       ssprfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix  is  symmetric  indefinite and
                 packed, provide error bounds and backward error estimates for
                 the solution

       sspsv  - compute the solution to a real system of linear equations  A*X
                 = B,  where A is an N-by-N symmetric matrix stored in  packed
                 format and X and B are N-by-NRHS matrices

       sspsvx  -  use  the diagonal pivoting factorization A = U*D*U**T or A =
                 L*D*L**T to compute the solution to a real system  of  linear
                 equations  A  *  X = B, where A is an N-by-N symmetric matrix
                 stored in packed format and X and B are N-by-NRHS matrices

       ssptrd - reduce a real symmetric matrix A stored in packed form to sym-
                 metric  tridiagonal form T by an orthogonal similarity trans-
                 formation

       ssptrf - compute the factorization of a real symmetric matrix A  stored
                 in  packed  format  using the Bunch-Kaufman diagonal pivoting
                 method

       ssptri - compute the inverse of a real symmetric indefinite matrix A in
                 packed  storage  using  the factorization A = U*D*U**T or A =
                 L*D*L**T computed by SSPTRF

       ssptrs - solve a system of linear equations A*X = B with a real symmet-
                 ric  matrix A stored in packed format using the factorization
                 A = U*D*U**T or A = L*D*L**T computed by SSPTRF

       sstebz - compute the eigenvalues of a symmetric tridiagonal matrix T

       sstedc - compute all eigenvalues and,  optionally,  eigenvectors  of  a
                 symmetric  tridiagonal  matrix  using  the divide and conquer
                 method

       sstegr - (a) Compute T-sigma_i = L_i D_i L_i^T, such that L_i D_i L_i^T
                 is a relatively robust representation

       sstein  -  compute  the  eigenvectors  of  a real symmetric tridiagonal
                 matrix  T  corresponding  to  specified  eigenvalues,   using
                 inverse iteration

       sstemr  - compute selected eigenvalues and, optionally, eigenvectors of
                 a real symmetric tridiagonal matrix T

       ssteqr - compute all eigenvalues and,  optionally,  eigenvectors  of  a
                 symmetric  tridiagonal  matrix  using  the  implicit QL or QR
                 method

       ssterf - compute all eigenvalues  of  a  symmetric  tridiagonal  matrix
                 using the Pal-Walker-Kahan variant of the QL or QR algorithm

       sstev - compute all eigenvalues and, optionally, eigenvectors of a real
                 symmetric tridiagonal matrix A

       sstevd - compute all eigenvalues and,  optionally,  eigenvectors  of  a
                 real symmetric tridiagonal matrix

       sstevr  - compute selected eigenvalues and, optionally, eigenvectors of
                 a real symmetric tridiagonal matrix T

       sstevx - compute selected eigenvalues and, optionally, eigenvectors  of
                 a real symmetric tridiagonal matrix A

       sstsv  - compute the solution to a system of linear equations A * X = B
                 where A is a symmetric tridiagonal matrix

       ssttrf - compute the factorization of a symmetric tridiagonal matrix  A
                 using the Bunch-Kaufman diagonal pivoting method

       ssttrs  - compute the solution to a real system of linear equations A*X
                 = B, where A is an N-by-N symmetric tridiagonal matrix and  X
                 and B are N-by-NRHS matrices

       sswap - Exchange vectors x and y

       ssycon  -  estimate  the  reciprocal  of  the  condition number (in the
                 1-norm) of a real symmetric matrix A using the  factorization
                 A = U*D*U**T or A = L*D*L**T computed by SSYTRF

       ssycon_rook  -  estimate the reciprocal of the condition number (in the
                 1-norm) of a real symmetric matrix using the factorization  A
                 = U*D*U**T or A = L*D*L**T computed by SSYTRF_ROOK

       ssyconv - convert A given by TRF into L and D and vice-versa

       ssyequb  -  compute  row  and column scalings intended to equilibrate a
                 symmetric matrix A  and  reduce  its  condition  number  with
                 respect to the two-norm

       ssyev - compute all eigenvalues and, optionally, eigenvectors of a real
                 symmetric matrix A

       ssyevd - compute all eigenvalues and,  optionally,  eigenvectors  of  a
                 real symmetric matrix A

       ssyevr  - compute selected eigenvalues and, optionally, eigenvectors of
                 a real symmetric tridiagonal matrix T

       ssyevx - compute selected eigenvalues and, optionally, eigenvectors  of
                 a real symmetric matrix A

       ssygs2  -  reduce a real symmetric-definite generalized eigenproblem to
                 standard form

       ssygst - reduce a real symmetric-definite generalized  eigenproblem  to
                 standard form

       ssygv  -  compute all the eigenvalues, and optionally, the eigenvectors
                 of a real generalized symmetric-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       ssygvd  - compute all the eigenvalues, and optionally, the eigenvectors
                 of a real generalized symmetric-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       ssygvx  - compute selected eigenvalues, and optionally, eigenvectors of
                 a real generalized symmetric-definite  eigenproblem,  of  the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       ssymm  -  perform  one  of the matrix-matrix operationsC := alpha*A*B +
                 beta*C or C := alpha*B*A + beta*C

       ssymv - perform the matrix-vector operationy := alpha*A*x + beta*y

       ssyr - perform the symmetric rank 1 operation   A := alpha*x*x' + A

       ssyr2 - perform the symmetric rank 2  operation    A  :=  alpha*x*y'  +
                 alpha*y*x' + A

       ssyr2k  -  perform  one  of  the  symmetric  rank  2k operations   C :=
                 alpha*A*B' +  alpha*B*A'  +  beta*C  or  C  :=  alpha*A'*B  +
                 alpha*B'*A + beta*C

       ssyrfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is symmetric indefinite,  provide
                 error bounds and backward error estimates for the solution

       ssyrfsx - improve the computed solution to a system of linear equations
                 when the coefficient matrix is symmetric indefinite,  provide
                 error bounds and backward error estimates for the solution

       ssyrk  -  perform  one  of  the  symmetric  rank  k  operations    C :=
                 alpha*A*A' + beta*C or C := alpha*A'*A + beta*C

       ssysv - compute the solution to a real system of linear equations   A*X
                 = B, where A is an N-by-N symmetric matrix and X and B are N-
                 by-NRHS matrices

       ssysv_rook - compute the solution to system of linear equations A*X = B
                 for  symmetric matrices. SSYTRF_ROOK is called to compute the
                 factorization of A

       ssysvx - use the diagonal pivoting factorization to compute  the  solu-
                 tion to a real system of linear equations A*X = B, where A is
                 an N-by-N symmetric matrix and X and B are N-by-NRHS matrices

       ssysvxx - compute the solution to real system of linear equations A*X =
                 B for symmetric matrices

       ssytd2 - reduce a real symmetric matrix A to symmetric tridiagonal form
                 T by an orthogonal similarity transformation

       ssytf2  -  computes  the  factorization  of a real symmetric indefinite
                 matrix, using the diagonal pivoting method  (unblocked  algo-
                 rithm).

       ssytf2_rook  - compute the factorization of a real symmetric indefinite
                 matrix using the bounded Bunch-Kaufman ("rook") diagonal piv-
                 oting method (unblocked algorithm)

       ssytrd - reduce a real symmetric matrix A to real symmetric tridiagonal
                 form T by an orthogonal similarity transformation

       ssytrf - compute the factorization of a real symmetric matrix  A  using
                 the Bunch-Kaufman diagonal pivoting method

       ssytrf_rook  -  compute  the  factorization  of a real symmetric matrix
                 using the bounded Bunch-Kaufman  ("rook")  diagonal  pivoting
                 method (blocked algorithm)

       ssytri  -  compute  the inverse of a real symmetric indefinite matrix A
                 using the factorization A = U*D*U**T or A = L*D*L**T computed
                 by SSYTRF

       ssytri2  -  compute the inverse of a REAL symmetric indefinite matrix A
                 using the factorization A = U*D*U**T or A = L*D*L**T computed
                 by SSYTRF

       ssytri2x  - compute the inverse of a real symmetric indefinite matrix A
                 using the factorization computed by SSYTRF

       ssytri_rook - compute the inverse of a real symmetric indefinite matrix
                 A  using  the factorization A = U*D*U**T or A = L*D*L**T com-
                 puted by SSYTRF_ROOK

       ssytrs - solve a system of linear equations A*X = B with a real symmet-
                 ric  matrix  A  using  the  factorization A = U*D*U**T or A =
                 L*D*L**T computed by SSYTRF

       ssytrs2 - solve a system of linear equations A*X = B with a  real  sym-
                 metric  matrix  A  using the factorization computed by SSYTRF
                 and converted by SSYCONV

       ssytrs_rook - solve a system of linear equations A*X = B  with  a  real
                 symmetric  matrix A using the factorization A = U*D*U**T or A
                 = L*D*L**T computed by SSYTRF_ROOK

       stbcon - estimate the reciprocal of the condition number of a  triangu-
                 lar band matrix A, in either the 1-norm or the infinity-norm

       stbmv  -  perform  one of the matrix-vector operationsx := A*x, or x :=
                 A'*x

       stbrfs - provide error bounds and  backward  error  estimates  for  the
                 solution  to  a  system of linear equations with a triangular
                 band coefficient matrix

       stbsv - solve one of the systems of equations A*x = b, or A'*x = b

       stbtrs - solve a triangular system of the form A*X = B or A**T*X  =  B,
                 where  A  is a triangular band matrix of order N, and B is an
                 N-by-NRHS matrix

       stfsm - solve a matrix equation (one operand is a triangular matrix  in
                 RFP format)

       stftri  -  compute  the  inverse of a triangular matrix A stored in RFP
                 format

       stfttp - copy a triangular matrix from the rectangular full packed for-
                 mat (TF) to the standard packed format (TP)

       stfttr - copy a triangular matrix from the rectangular full packed for-
                 mat (TF) to the standard full format (TR)

       stgevc -  compute the right and/or left generalized eigenvectors  of  a
                 pair  of  real  upper  triangular  matrices  (A,B)  that  was
                 obtained from the generalized Schur factorization of an orig-
                 inal pair of real nonsymmetric matrices (AO,BO)

       stgexc  -  reorder the generalized Schur decomposition of a real matrix
                 pair using an orthogonal or unitary  equivalence  transforma-
                 tion

       stgsen  -  reorder  the  generalized real Schur decomposition of a real
                 matrix pair (A, B), so that a selected cluster of eigenvalues
                 appears  in  the  leading diagonal blocks of the upper quasi-
                 triangular matrix A and the upper triangular B

       stgsja - compute the generalized singular value decomposition (GSVD) of
                 two real upper triangular (or trapezoidal) matrices A and B

       stgsna  - estimate reciprocal condition numbers for specified eigenval-
                 ues and/or eigenvectors of a matrix pair (A, B)  in  general-
                 ized  real  Schur  canonical  form  (or  of  any  matrix pair
                 (Q*A*Z', Q*B*Z') with orthogonal matrices Q and Z,  where  Z'
                 denotes the transpose of Z

       stgsyl - solve the generalized Sylvester equation

       stpcon  -  estimate  the reciprocal of the condition number of a packed
                 triangular matrix A, in either the 1-norm  or  the  infinity-
                 norm

       stpmqrt - apply a real orthogonal matrix Q obtained from a "triangular-
                 pentagonal" real block reflector H to a general  real  matrix
                 C, which consists of two blocks

       stpmv  -  perform one of the matrix-vector operations x := A*x, or x :=
                 A'*x

       stpqrt - compute a blocked QR factorization of a real  "triangular-pen-
                 tagonal"  matrix C, which is composed of a triangular block A
                 and pentagonal block B, using the compact  WY  representation
                 for Q

       stpqrt2  - compute a QR factorization of a real or complex "triangular-
                 pentagonal" matrix, which is composed of a  triangular  block
                 and  a  pentagonal block, using the compact WY representation
                 for Q

       stprfs - provide error bounds and  backward  error  estimates  for  the
                 solution  to  a  system of linear equations with a triangular
                 packed coefficient matrix

       stpsv - solve one of the systems of equations A*x = b, or A'*x = b

       stptri - compute the inverse of a real upper or lower triangular matrix
                 A stored in packed format

       stptrs  -  solve a triangular system of the form A*X = B or A**T*X = B,
                 where A is a triangular matrix of order N  stored  in  packed
                 format, and B is an N-by-NRHS matrix

       stpttf  - copy a triangular matrix from the standard packed format (TP)
                 to the rectangular full packed format (TF)

       stpttr - copy a triangular matrix from the standard packed format  (TP)
                 to the standard full format (TR)

       strans - transpose and scale source matrix

       strcon  - estimate the reciprocal of the condition number of a triangu-
                 lar matrix A, in either the 1-norm or the infinity-norm

       strevc - compute some or all of the right and/or left eigenvectors of a
                 real upper quasi-triangular matrix T

       strexc  -  reorder  the  real  Schur factorization of a real matrix A =
                 Q*T*Q**T, so that the diagonal block of T with row index IFST
                 is moved to row ILST

       strmm  -  perform  one  of the matrix-matrix operationsB := alpha*op( A
                 )*B, or B := alpha*B*op( A )

       strmv - perform one of the matrix-vector operations x := A*x, or  x  :=
                 A'*x

       strrfs  -  provide  error  bounds  and backward error estimates for the
                 solution to a system of linear equations  with  a  triangular
                 coefficient matrix

       strsen  -  reorder  the  real  Schur factorization of a real matrix A =
                 Q*T*Q**T, so that a selected cluster of  eigenvalues  appears
                 in  the leading diagonal blocks of the upper quasi-triangular
                 matrix T,

       strsm - solve one of the matrix equations   op( A  )*X  =  alpha*B,  or
                 X*op( A ) = alpha*B

       strsna  - estimate reciprocal condition numbers for specified eigenval-
                 ues and/or right eigenvectors of a real upper  quasi-triangu-
                 lar matrix T (or of any matrix Q*T*Q**T with Q orthogonal)

       strsv - solve one of the systems of equations A*x = b, or A'*x = b

       strsyl - solve the real Sylvester matrix equation

       strti2 - compute the inverse of a real upper or lower triangular matrix

       strtri - compute the inverse of a real upper or lower triangular matrix
                 A

       strtrs  - solve a triangular system of the form  A*X = B or A**T*X = B,
                 where A is a triangular matrix of order N, and B is an  N-by-
                 NRHS matrix

       strttf - copy a triangular matrix from the standard full format (TR) to
                 the rectangular full packed format (TF)

       strttp - copy a triangular matrix from the standard full format (TR) to
                 the standard packed format (TP)

       stzrqf - routine is deprecated and has been replaced by routine STZRZF

       stzrzf  - reduce the M-by-N ( M<=N ) real upper trapezoidal matrix A to
                 upper triangular form by means of orthogonal transformations

       sunperf - Introduction to Oracle  Solaris  Studio  Performance  Library
                 functions and subroutines

       sunperf_version - gets library information

       svbrmm - variable block sparse row format matrix-matrix multiply

       svbrsm - variable block sparse row format triangular solve

       swiener - perform Wiener deconvolution of two signals

       use_threads - Sets the number of threads to use for subsequent parallel
                 regions

       using_threads - In a parallel environment,  if  called  from  a  serial
                 region  of the master thread it returns the number of threads
                 available   for    execution    (determined    by    function
                 OMP_GET_NUM_THREADS).  Else, if it is called from a thread in
                 the team executing  the  parallel  region  it  returns  a  1.
                 USING_THREADS subroutine

       vcfftb  -  compute  a  periodic sequence from its Fourier coefficients.
                 The VCFFT operations are normalized, so a call of VCFFTF fol-
                 lowed  by a call of VCFFTB will return the original sequence.

       vcfftf - compute the Fourier coefficients of a periodic sequence.   The
                 VCFFT operations are normalized, so a call of VCFFTF followed
                 by a call of VCFFTB will return the original sequence.

       vcffti - initialize the array WSAVE, which is used in both  VCFFTF  and
                 VCFFTB.

       vcosqb - synthesize a Fourier sequence from its representation in terms
                 of a cosine series with odd wave numbers.  The  VCOSQ  opera-
                 tions  are normalized, so a call of VCOSQF followed by a call
                 of VCOSQB will return the original sequence.

       vcosqf - compute the Fourier coefficients in a cosine series  represen-
                 tation  with  only odd wave numbers. The VCOSQ operations are
                 normalized, so a call of VCOSQF followed by a call of  VCOSQB
                 will return the original sequence.

       vcosqi  -  initialize the array WSAVE, which is used in both VCOSQF and
                 VCOSQB.

       vcost - compute the  discrete  Fourier  cosine  transform  of  an  even
                 sequence.   The  VCOST  transform is normalized, so a call of
                 VCOST followed by a call of VCOST will  return  the  original
                 sequence.

       vcosti - initialize the array WSAVE, which is used in VCOST.

       vdcosqb  -  synthesize  a  Fourier  sequence from its representation in
                 terms of a cosine series with odd  wave  numbers.  The  VCOSQ
                 operations  are normalized, so a call of VCOSQF followed by a
                 call of VCOSQB will return the original sequence.

       vdcosqf - compute the Fourier coefficients in a cosine series represen-
                 tation  with  only odd wave numbers. The VCOSQ operations are
                 normalized, so a call of VCOSQF followed by a call of  VCOSQB
                 will return the original sequence.

       vdcosqi  - initialize the array WSAVE, which is used in both VCOSQF and
                 VCOSQB.

       vdcost - compute the discrete  Fourier  cosine  transform  of  an  even
                 sequence.   The  VCOST  transform is normalized, so a call of
                 VCOST followed by a call of VCOST will  return  the  original
                 sequence.

       vdcosti - initialize the array WSAVE, which is used in VCOST.

       vdfftb  -  compute  a  periodic sequence from its Fourier coefficients.
                 The VDFFT operations are normalized, so a call of VDFFTF fol-
                 lowed  by a call of VDFFTB will return the original sequence.

       vdfftf - compute the Fourier coefficients of a periodic sequence.   The
                 VDFFT operations are normalized, so a call of VDFFTF followed
                 by a call of VDFFTB will return the original sequence.

       vdffti - initialize the array WSAVE, which is used in both  VRFFTF  and
                 VRFFTB.

       vdsinqb  -  synthesize  a  Fourier  sequence from its representation in
                 terms of a sine series with  odd  wave  numbers.   The  VSINQ
                 operations  are normalized, so a call of VSINQF followed by a
                 call of VSINQB will return the original sequence.

       vdsinqf - compute the Fourier coefficients in a sine series representa-
                 tion with only odd wave numbers.The VSINQ operations are nor-
                 malized, so a call of VSINQF followed by  a  call  of  VSINQB
                 will return the original sequence.

       vdsinqi  - initialize the array WSAVE, which is used in both VSINQF and
                 VSINQB.

       vdsint -  compute  the  discrete  Fourier  sine  transform  of  an  odd
                 sequence.  The  VSINT transforms are unnormalized inverses of
                 themselves, so a call of VSINT followed by  another  call  of
                 VSINT  will  multiply  the  input sequence by 2 * (N+1).  The
                 VSINT transforms are normalized, so a call of VSINT  followed
                 by a call of VSINT will return the original sequence.

       vdsinti  -  initialize  the  array  WSAVE,  which is used in subroutine
                 VSINT.

       vrfftb - compute a periodic sequence  from  its  Fourier  coefficients.
                 The VRFFT operations are normalized, so a call of VRFFTF fol-
                 lowed by a call of VRFFTB will return the original  sequence.

       vrfftf  - compute the Fourier coefficients of a periodic sequence.  The
                 VRFFT operations are normalized, so a call of VRFFTF followed
                 by a call of VRFFTB will return the original sequence.

       vrffti  -  initialize the array WSAVE, which is used in both VRFFTF and
                 VRFFTB.

       vsinqb - synthesize a Fourier sequence from its representation in terms
                 of a sine series with odd wave numbers.  The VSINQ operations
                 are normalized, so a call of VSINQF followed  by  a  call  of
                 VSINQB will return the original sequence.

       vsinqf  - compute the Fourier coefficients in a sine series representa-
                 tion with only odd wave numbers.The VSINQ operations are nor-
                 malized,  so  a  call  of VSINQF followed by a call of VSINQB
                 will return the original sequence.

       vsinqi - initialize the array WSAVE, which is used in both  VSINQF  and
                 VSINQB.

       vsint - compute the discrete Fourier sine transform of an odd sequence.
                 The VSINT transforms are unnormalized inverses of themselves,
                 so  a  call  of  VSINT followed by another call of VSINT will
                 multiply the input sequence by 2 * (N+1).  The  VSINT  trans-
                 forms  are  normalized, so a call of VSINT followed by a call
                 of VSINT will return the original sequence.

       vsinti - initialize the array WSAVE, which is used in subroutine VSINT.

       vzfftb  -  compute  a  periodic sequence from its Fourier coefficients.
                 The VZFFT operations are normalized, so a call of VZFFTF fol-
                 lowed  by a call of VZFFTB will return the original sequence.

       vzfftf - compute the Fourier coefficients of a periodic sequence.   The
                 VZFFT operations are normalized, so a call of VZFFTF followed
                 by a call of VZFFTB will return the original sequence.

       vzffti - initialize the array WSAVE, which is used in both  VZFFTF  and
                 VZFFTB.

       zCopy_CompCol_Matrix  -  A  utility  C  function  in the serial SuperLU
                 solver that copies one SuperMatrix into another.

       zCreate_CompCol_Matrix - A utility C function  in  the  serial  SuperLU
                 solver that creates a SuperMatrix in compressed sparse column
                 format (also known as the Harwell-Boeing format).

       zCreate_CompRow_Matrix - A utility C function  in  the  serial  SuperLU
                 solver  that  creates  a SuperMatrix in compressed sparse row
                 format.

       zCreate_Dense_Matrix - A utility  C  function  in  the  serial  SuperLU
                 solver that creates a SuperMatrix in dense format.

       zCreate_SuperNode_Matrix  -  A utility C function in the serial SuperLU
                 solver that creates a SuperMatrix in supernodal format.

       zPrintPerf - A utility function of the SuperLU solver that prints  sta-
                 tistics collected by the computational routines.

       zQuerySpace  - A inquiry function that provides information on the mem-
                 ory statistics of the SuperLU solver.

       zaxpy - compute y := alpha * x + y

       zaxpyi - Compute y := alpha * x + y

       zbbcsd - compute the CS decomposition of a unitary matrix  in  bidiago-
                 nal-block form

       zbcomm - block coordinate matrix-matrix multiply

       zbdimm - block diagonal format matrix-matrix multiply

       zbdism -  block diagonal format triangular solve

       zbdsqr - compute the singular value decomposition (SVD) of a real N-by-
                 N (upper or lower) bidiagonal matrix B

       zbelmm - block Ellpack format matrix-matrix multiply

       zbelsm - block Ellpack format triangular solve

       zbscmm - block sparse column matrix-matrix multiply

       zbscsm - block sparse column format triangular solve

       zbsrmm - block sparse row format matrix-matrix multiply

       zbsrsm - block sparse row format triangular solve

       zcgesv - computes the solution to a complex system of linear  equations
                 A * X = B

       zcnvcor - compute the convolution or correlation of complex vectors

       zcnvcor2 - compute the convolution or correlation of complex matrices

       zcoomm - coordinate matrix-matrix multiply

       zcopy - copy x to y

       zcposv  - computes the solution to system of linear equations A * X = B
                 for PO matrices

       zcscmm - compressed sparse column format matrix-matrix multiply

       zcscsm - compressed sparse column format triangular solve

       zcsrmm - compressed sparse row format matrix-matrix multiply

       zcsrsm - compressed sparse row format triangular solve

       zdiamm - diagonal format matrix-matrix multiply.

       zdiasm - diagonal format triangular solve

       zdotc - compute the dot product of two vectors conjg(x) and y.

       zdotci - Compute the complex conjugated indexed dot product.

       zdotu - compute the dot product of two vectors x and y.

       zdotui - Compute the complex unconjugated indexed dot product.

       zdrot - Apply a plane rotation.

       zdscal - Compute y := alpha * y

       zellmm - Ellpack format matrix-matrix multiply

       zellsm - Ellpack format triangular solve

       zfft2b - compute a periodic sequence  from  its  Fourier  coefficients.
                 The FFT operations are unnormalized, so a call of ZFFT2F fol-
                 lowed by a call of ZFFT2B will multiply the input sequence by
                 M*N.

       zfft2f  - compute the Fourier coefficients of a periodic sequence.  The
                 FFT operations are unnormalized, so a call of ZFFT2F followed
                 by  a call of ZFFT2B will multiply the input sequence by M*N.

       zfft2i - initialize the array WSAVE, which is used in both the  forward
                 and backward transforms.

       zfft3b  -  compute  a  periodic sequence from its Fourier coefficients.
                 The FFT operations are unnormalized, so a call of ZFFT3F fol-
                 lowed by a call of ZFFT3B will multiply the input sequence by
                 M*N*K.

       zfft3f - compute the Fourier coefficients of a periodic sequence.   The
                 FFT operations are unnormalized, so a call of ZFFT3F followed
                 by a call of ZFFT3B  will  multiply  the  input  sequence  by
                 M*N*K.

       zfft3i  -  initialize the array WSAVE, which is used in both ZFFT3F and
                 ZFFT3B.

       zfftb - compute a periodic sequence from its Fourier coefficients.  The
                 FFT  operations are unnormalized, so a call of ZFFTF followed
                 by a call of ZFFTB will multiply the input sequence by N.

       zfftd - initialize the trigonometric weight and factor tables  or  com-
                 pute  the  inverse Fast Fourier Transform of a double complex
                 sequence.

       zfftd2 - initialize the trigonometric weight and factor tables or  com-
                 pute  the two-dimensional inverse Fast Fourier Transform of a
                 two-dimensional double complex array.

       zfftd3 - initialize the trigonometric weight and factor tables or  com-
                 pute  the three-dimensional inverse Fast Fourier Transform of
                 a three-dimensional double complex array.

       zfftdm - initialize the trigonometric weight and factor tables or  com-
                 pute  the one-dimensional inverse Fast Fourier Transform of a
                 set of double complex data sequences stored in  a  two-dimen-
                 sional array.

       zfftf  -  compute the Fourier coefficients of a periodic sequence.  The
                 FFT operations are unnormalized, so a call of ZFFTF  followed
                 by a call of ZFFTB will multiply the input sequence by N.

       zffti  -  initialize  the  array WSAVE, which is used in both ZFFTF and
                 ZFFTB.

       zfftopt - compute the length of the closest fast FFT

       zfftz - initialize the trigonometric weight and factor tables  or  com-
                 pute  the  Fast  Fourier  transform (forward or inverse) of a
                 double complex sequence.

       zfftz2 - initialize the trigonometric weight and factor tables or  com-
                 pute  the  two-dimensional Fast Fourier Transform (forward or
                 inverse) of a two-dimensional double complex array.

       zfftz3 - initialize the trigonometric weight and factor tables or  com-
                 pute the three-dimensional Fast Fourier Transform (forward or
                 inverse) of a three-dimensional double complex array.

       zfftzm - initialize the trigonometric weight and factor tables or  com-
                 pute  the  one-dimensional Fast Fourier Transform (forward or
                 inverse) of a set of data sequences stored  in  a  two-dimen-
                 sional double complex array.

       zgbbrd  -  reduce  a complex general m-by-n band matrix A to real upper
                 bidiagonal form B by a unitary transformation

       zgbcon - estimate the reciprocal of the condition number of  a  complex
                 general  band matrix A, in either the 1-norm or the infinity-
                 norm, using the LU factorization computed by ZGBTRF

       zgbequ - compute row and column scalings intended to equilibrate an  M-
                 by-N band matrix A and reduce its condition number

       zgbequb - compute row and column scalings intended to equilibrate an M-
                 by-N matrix A and reduce its condition number

       zgbmv - perform one of the matrix-vector  operationsy  :=  alpha*A*x  +
                 beta*y,  or  y := alpha*A'*x + beta*y, or   y := alpha*conjg(
                 A' )*x + beta*y

       zgbrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix  is  banded, and provide error
                 bounds and backward error estimates for the solution

       zgbrfsx - improve the computed solution to a system of linear equations
                 and provide error bounds and backward error estimates for the
                 solution

       zgbsv - compute the solution to a complex system of linear equations  A
                 * X = B, where A is a band matrix of order N with KL subdiag-
                 onals and KU superdiagonals, and X and B are N-by-NRHS matri-
                 ces

       zgbsvx  - use the LU factorization to compute the solution to a complex
                 system of linear  equations  A*X=B,  A**T*X=B,  or  A**H*X=B,
                 where A is a band matrix

       zgbsvxx  - compute the solution to system of linear equations A * X = B
                 for general band matrices

       zgbtf2 - compute an LU factorization of a complex m-by-n band matrix  A
                 using partial pivoting with row interchanges

       zgbtrf  - compute an LU factorization of a complex m-by-n band matrix A
                 using partial pivoting with row interchanges

       zgbtrs - solve  a  system  of  linear  equations  A*X=B,  A**T*X=B,  or
                 A**H*X=B with a general band matrix A using the LU factoriza-
                 tion computed by ZGBTRF

       zgebak - form the right or  left  eigenvectors  of  a  complex  general
                 matrix  by  backward transformation on the computed eigenvec-
                 tors of the balanced matrix output by ZGEBAL

       zgebal - balance a general complex matrix A

       zgebd2 - reduce a general matrix to bidiagonal form using an  unblocked
                 algorithm

       zgebrd  -  reduce  a  general complex M-by-N matrix A to upper or lower
                 bidiagonal form B by a unitary transformation

       zgecon - estimate the reciprocal of the condition number of  a  general
                 complex  matrix A, in either the 1-norm or the infinity-norm,
                 using the LU factorization computed by ZGETRF

       zgeequ - compute row and column scalings intended to equilibrate an  M-
                 by-N matrix A and reduce its condition number

       zgeequb  -  computes row and column scalings intended to equilibrate an
                 M-by-N matrix A and redu ce its condition number

       zgees - compute for an N-by-N complex nonsymmetric matrix A, the eigen-
                 values,  the  Schur  form  T,  and, optionally, the matrix of
                 Schur vectors Z

       zgeesx - compute for an N-by-N complex nonsymmetric matrix A,  the  ei-
                 genvalues,  the  Schur form T, and, optionally, the matrix of
                 Schur vectors Z

       zgeev - compute for an N-by-N complex nonsymmetric matrix A, the eigen-
                 values and, optionally, the left and/or right eigenvectors

       zgeevx  -  compute for an N-by-N complex nonsymmetric matrix A, the ei-
                 genvalues and, optionally, the left and/or right eigenvectors

       zgegs - routine is deprecated and has been replaced by routine ZGGES

       zgegv - routine is deprecated and has been replaced by routine ZGGEV

       zgehd2  - reduce a general square matrix to upper Hessenberg form using
                 an unblocked algorithm

       zgehrd - reduce a complex general matrix A to upper Hessenberg  form  H
                 by a unitary similarity transformation

       zgelq2  -  compute the LQ factorization of a general rectangular matrix
                 using an unblocked algorithm

       zgelqf - compute an LQ factorization of a complex M-by-N matrix A

       zgels - solve overdetermined or underdetermined complex linear  systems
                 involving  an  M-by-N  matrix  A, or its conjugate-transpose,
                 using a QR or LQ factorization of A

       zgelsd - compute the minimum-norm  solution  to  a  real  linear  least
                 squares problem

       zgelss  -  compute  the minimum norm solution to a complex linear least
                 squares problem

       zgelsx - routine is deprecated and has been replaced by routine ZGELSY

       zgelsy - compute the minimum-norm solution to a  complex  linear  least
                 squares problem

       zgemm  -  perform  one of the matrix-matrix operations C := alpha*op( A
                 )*op( B ) + beta*C

       zgemqrt - overwrite the general complex M-by-N matrix C with Q*C,  C*Q,
                 Q**H* C, or C*Q**H depending on values of SIDE and TRANS

       zgemv  -  perform  one  of the matrix-vector operationsy := alpha*A*x +
                 beta*y, or y := alpha*A'*x + beta*y, or   y  :=  alpha*conjg(
                 A' )*x + beta*y

       zgeql2  -  compute the QL factorization of a general rectangular matrix
                 using an unblocked algorithm

       zgeqlf - compute a QL factorization of a complex M-by-N matrix A

       zgeqp3 - compute a QR factorization with column pivoting of a matrix A

       zgeqpf - routine is deprecated and has been replaced by routine ZGEQP3

       zgeqr2 - computes the QR factorization of a general rectangular  matrix
                 using an unblocked algorithm.

       zgeqr2p - computes the QR factorization of a general rectangular matrix
                 with non-negative diagonal elements using an unblocked  algo-
                 rithm.

       zgeqrf - compute a QR factorization of a complex M-by-N matrix A

       zgeqrfp  - compute a QR factorization of a complex M-by-N matrix A: A =
                 Q * R

       zgeqrt - compute a blocked QR factorization of a complex M-by-N  matrix
                 A using the compact WY representation of Q

       zgeqrt2  - compute a QR factorization of a general complex matrix using
                 the compact WY representation of Q

       zgeqrt3 - recursively computes a QR factorization of a general  complex
                 matrix using the compact WY representation of Q

       zgerc - perform the rank 1 operation A := alpha*x*conjg( y' ) + A

       zgerfs  - improve the computed solution to a system of linear equations
                 and provide error bounds and backward error estimates for the
                 solution

       zgerfsx - improve the computed solution to a system of linear equations
                 and provide error bounds and backward error estimates for the
                 solution

       zgerq2  - computes the RQ factorization of a general rectangular matrix
                 using an unblocked algorithm

       zgerqf - compute an RQ factorization of a complex M-by-N matrix A

       zgeru - perform the rank 1 operation A := alpha*x*y' + A

       zgesdd - compute the singular value decomposition (SVD) of a complex M-
                 by-N  matrix  A,  optionally  computing the left and/or right
                 singular vectors, by using divide-and-conquer method

       zgesv - compute the solution to a complex system  of  linear  equations
                 A*X=B,  where A is an N-by-N matrix and X and B are N-by-NRHS
                 matrices

       zgesvd - compute the singular value decomposition (SVD) of a complex M-
                 by-N  matrix  A,  optionally  computing the left and/or right
                 singular vectors

       zgesvx - use the LU factorization to compute the solution to a  complex
                 system  of linear equations  A*X=B, where A is an N-by-N gen-
                 eral matrix

       zgesvxx - compute the solution to system of linear equations A*X=B  for
                 general matrices

       zgetf2 - compute an LU factorization of a general m-by-n matrix A using
                 partial pivoting with row interchanges

       zgetrf - compute an LU factorization of a general M-by-N matrix A using
                 partial pivoting with row interchanges

       zgetri  -  compute  the  inverse of a matrix using the LU factorization
                 computed by ZGETRF

       zgetrs - solve a system of linear equations  A * X = B, A**T * X  =  B,
                 or  A**H  * X = B with a general N-by-N matrix A using the LU
                 factorization computed by ZGETRF

       zggbak - form the right or left eigenvectors of a  complex  generalized
                 eigenvalue  problem A*x = lambda*B*x, by backward transforma-
                 tion on the computed eigenvectors of  the  balanced  pair  of
                 matrices output by ZGGBAL

       zggbal - balance a pair of general complex matrices (A,B)

       zgges  -  compute  for  a  pair of N-by-N complex nonsymmetric matrices
                 (A,B), the generalized eigenvalues, the  generalized  complex
                 Schur  form  (S,  T),  and optionally left and/or right Schur
                 vectors (VSL and VSR)

       zggesx - compute for a pair of  N-by-N  complex  nonsymmetric  matrices
                 (A,B),  the  generalized  eigenvalues, the complex Schur form
                 (S,T), and, optionally, the left  and/or  right  matrices  of
                 Schur vectors

       zggev  -  compute  for  a  pair of N-by-N complex nonsymmetric matrices
                 (A,B), the generalized eigenvalues, and optionally, the  left
                 and/or right generalized eigenvectors

       zggevx  -  compute  for  a pair of N-by-N complex nonsymmetric matrices
                 (A,B) the generalized eigenvalues, and, optionally, the  left
                 and/or right generalized eigenvectors

       zggglm - solve a general Gauss-Markov linear model (GLM) problem

       zgghrd  -  reduce a pair of complex matrices (A,B) to generalized upper
                 Hessenberg form using unitary transformations, where A  is  a
                 general matrix and B is upper triangular

       zgglse  -  solve  the  linear  equality-constrained least squares (LSE)
                 problem

       zggqrf - compute a generalized QR factorization of an N-by-M  matrix  A
                 and an N-by-P matrix B.

       zggrqf  -  compute a generalized RQ factorization of an M-by-N matrix A
                 and a P-by-N matrix B

       zggsvd - compute the generalized singular value decomposition (GSVD) of
                 an M-by-N complex matrix A and P-by-N complex matrix B

       zggsvp - compute unitary matrices

       zgscon  - estimates the reciprocal of the condition number of a general
                 real matrix A, in either the  1-norm  or  the  infinity-norm,
                 using  the  LU  factorization  computed  by  SuperLU  routine
                 sgstrf.

       zgsequ - computes row and column scalings intended to equilibrate an M-
                 by-N sparse matrix A and reduce its condition number.

       zgsrfs - improves the computed solution to a system of linear equations
                 and provides error bounds and backward  error  estimates  for
                 the solution.  It is a SuperLU routine.

       zgssco - General sparse solver condition number estimate.

       zgssda - Deallocate working storage for the general sparse solver.

       zgssfa - General sparse solver numeric factorization.

       zgssfs - General sparse solver one call interface.

       zgssin - Initialize the general sparse solver.

       zgssor - General sparse solver ordering and symbolic factorization.

       zgssps - Print general sparse solver statics.

       zgssrp - Return permutation used by the general sparse solver.

       zgsssl - Solve routine for the general sparse solver.

       zgssuo  -  Provide  general sparse solvers SPSOLVE and SuperLU  a user-
                 supplied permutation for ordering.

       zgssv - solves a system of linear equations A*X=B using the LU  factor-
                 ization from sgstrf.

       zgssvx  -  solves the system of linear equations A*X=B or A'*X=B, using
                 the LU factorization from sgstrf(). Error bounds on the solu-
                 tion and a condition estimate are also provided.

       zgstrf - computes an LU factorization of a general sparse m-by-n matrix
                 A using partial pivoting with row interchanges.

       zgstrs - solves a system of linear equations A*X=B  or  A'*X=B  with  A
                 sparse  and  B  dense, using the LU factorization computed by
                 sgstrf.

       zgtcon - estimate the reciprocal of the condition number of  a  complex
                 tridiagonal  matrix  A using the LU factorization as computed
                 by ZGTTRF

       zgthr - Gathers specified elements from y into x.

       zgthrz - Gather and zero.

       zgtrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix  is tridiagonal, provide error
                 bounds and backward error estimates for the solution

       zgtsv - solve the equation A*X = B, where A is  an  N-by-N  tridiagonal
                 matrix, by Gaussian elimination with partial pivoting

       zgtsvx  - use the LU factorization to compute the solution to a complex
                 system of linear  equations  A*X=B,  A**T*X=B,  or  A**H*X=B,
                 where A is a tridiagonal matrix of order N and X and B are N-
                 by-NRHS matrices

       zgttrf - compute an LU factorization of a complex tridiagonal matrix  A
                 using elimination with partial pivoting and row interchanges

       zgttrs - solve one of the systems of equations A*X=B, A**T*X=B, or A**H
                 *X=B, with a tridiagonal matrix A using the LU  factorization
                 computed by ZGTTRF

       zgtts2  -  solve a system of linear equations with a tridiagonal matrix
                 using the LU factorization computed by zgttrf

       zhbev - compute all the eigenvalues and, optionally, eigenvectors of  a
                 complex Hermitian band matrix A

       zhbevd - compute all the eigenvalues and, optionally, eigenvectors of a
                 complex Hermitian band matrix A

       zhbevx - compute selected eigenvalues and, optionally, eigenvectors  of
                 a complex Hermitian band matrix A

       zhbgst  - reduce a complex Hermitian-definite banded generalized eigen-
                 problem A*x=lambda*B*x to standard  form  C*y=lambda*y,  such
                 that C has the same bandwidth as A

       zhbgv  -  compute all the eigenvalues, and optionally, the eigenvectors
                 of a complex generalized Hermitian-definite banded eigenprob-
                 lem, of the form A*x=(lambda)*B*x

       zhbgvd  - compute all the eigenvalues, and optionally, the eigenvectors
                 of a complex generalized Hermitian-definite banded eigenprob-
                 lem, of the form A*x=(lambda)*B*x

       zhbgvx  - compute all the eigenvalues, and optionally, the eigenvectors
                 of a complex generalized Hermitian-definite banded eigenprob-
                 lem, of the form A*x=(lambda)*B*x

       zhbmv - perform the matrix-vector operationy := alpha*A*x + beta*y

       zhbtrd  -  reduce  a  complex Hermitian band matrix A to real symmetric
                 tridiagonal form T by a unitary similarity transformation

       zhecon - estimate the reciprocal of the condition number of  a  complex
                 Hermitian  matrix A using the factorization A = U*D*U**H or A
                 = L*D*L**H computed by ZHETRF

       zhecon_rook - estimate the reciprocal of the condition number for  Her-
                 mitian  matrices using factorization obtained with one of the
                 bounded diagonal pivoting methods (max 2 interchanges)

       zheequb - compute row and column scalings  intended  to  equilibrate  a
                 Hermitian  matrix  A  and  reduce  its condition number (with
                 respect to the two-norm)

       zheev - compute all eigenvalues and, optionally, eigenvectors of a com-
                 plex Hermitian matrix A

       zheevd  -  compute  all  eigenvalues and, optionally, eigenvectors of a
                 complex Hermitian matrix A

       zheevr - compute selected eigenvalues and, optionally, eigenvectors  of
                 a complex Hermitian tridiagonal matrix T

       zheevx  - compute selected eigenvalues and, optionally, eigenvectors of
                 a complex Hermitian matrix A

       zhegs2 - reduce a complex Hermitian-definite  generalized  eigenproblem
                 to standard form

       zhegst  -  reduce a complex Hermitian-definite generalized eigenproblem
                 to standard form

       zhegv - compute all the eigenvalues, and optionally,  the  eigenvectors
                 of  a complex generalized Hermitian-definite eigenproblem, of
                 the    form     A*x=(lambda)*B*x,     A*Bx=(lambda)*x,     or
                 B*A*x=(lambda)*x

       zhegvd  - compute all the eigenvalues, and optionally, the eigenvectors
                 of a complex generalized Hermitian-definite eigenproblem,  of
                 the     form     A*x=(lambda)*B*x,     A*Bx=(lambda)*x,    or
                 B*A*x=(lambda)*x

       zhegvx - compute selected eigenvalues, and optionally, eigenvectors  of
                 a complex generalized Hermitian-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       zhemm - perform one of the matrix-matrix operations C  :=  alpha*A*B  +
                 beta*C or C := alpha*B*A + beta*C

       zhemv - perform the matrix-vector operationy := alpha*A*x + beta*y

       zher  - perform the hermitian rank 1 operation   A := alpha*x*conjg( x'
                 ) + A

       zher2 - perform the hermitian rank 2 operation   A := alpha*x*conjg( y'
                 ) + conjg( alpha )*y*conjg( x' ) + A

       zher2k  -  perform  one  of  the  Hermitian  rank  2k operations   C :=
                 alpha*A*conjg( B' ) + conjg( alpha )*B*conjg( A' )  +  beta*C
                 or  C := alpha*conjg( A' )*B + conjg( alpha )*conjg( B' )*A +
                 beta*C

       zherfs - improve the computed solution to a system of linear  equations
                 when the coefficient matrix is Hermitian indefinite, and pro-
                 vide error bounds and backward error estimates for the  solu-
                 tion

       zherfsx - improve the computed solution to a system of linear equations
                 when the coefficient matrix is Hermitian indefinite, and pro-
                 vide  error bounds and backward error estimates for the solu-
                 tion

       zherk -  perform  one  of  the  Hermitian  rank  k  operations    C  :=
                 alpha*A*conjg(  A'  )  + beta*C or C := alpha*conjg( A' )*A +
                 beta*C

       zhesv - compute the solution to a complex system  of  linear  equations
                 A*X=B,  where A is an N-by-N Hermitian matrix and X and B are
                 N-by-NRHS matrices

       zhesv_rook - compute the solution to a system of linear equations A*X=B
                 for   Hermitian  matrices  using  the  bounded  Bunch-Kaufman
                 ("rook") diagonal pivoting method

       zhesvx - use the diagonal pivoting factorization to compute  the  solu-
                 tion to a complex system of linear equations A*X = B, where A
                 is an N-by-N Hermitian matrix  and  X  and  B  are  N-by-NRHS
                 matrices

       zhesvxx  -  compute  the solution to system of linear equations A*X = B
                 for Hermitian matrices

       zhetd2 - reduce a Hermitian matrix to real symmetric  tridiagonal  form
                 by an unitary similarity transformation (unblocked algorithm)

       zhetf2 - compute the factorization of a complex Hermitian matrix, using
                 the  diagonal  pivoting  method (unblocked algorithm, calling
                 Level 2 BLAS)

       zhetf2_rook - compute the factorization of a complex Hermitian  indefi-
                 nite matrix using the bounded Bunch-Kaufman ("rook") diagonal
                 pivoting method (unblocked algorithm)

       zhetrd - reduce a complex Hermitian matrix A to real symmetric tridiag-
                 onal form T by a unitary similarity transformation

       zhetrf  -  compute  the  factorization  of a complex Hermitian matrix A
                 using the Bunch-Kaufman diagonal pivoting method

       zhetrf_rook - compute the factorization of a complex Hermitian  indefi-
                 nite matrix using the bounded Bunch-Kaufman ("rook") diagonal
                 pivoting method (blocked algorithm, calling Level 3 BLAS)

       zhetri - compute the inverse of a complex Hermitian indefinite matrix A
                 using the factorization A = U*D*U**H or A = L*D*L**H computed
                 by ZHETRF

       zhetri2 - compute the inverse  of  a  COMPLEX*16  Hermitian  indefinite
                 matrix  A  using  the  factorization A=U*D*U**T or A=L*D*L**T
                 computed by ZHETRF

       zhetri2x - compute the inverse of  a  COMPLEX*16  Hermitian  indefinite
                 matrix A using the factorization A = U*D*U**H or A = L*D*L**H
                 computed by ZHETRF

       zhetri_rook - compute the inverse of a Hermitian matrix using the  fac-
                 torization  obtained  with the bounded Bunch-Kaufman ("rook")
                 diagonal pivoting method

       zhetrs - solve a system of linear equations A*X = B with a complex Her-
                 mitian  matrix  A using the factorization A = U*D*U**H or A =
                 L*D*L**H computed by ZHETRF

       zhetrs2 - solve a system of linear equations A*X=B with a complex  Her-
                 mitian   matrix  A  using  the  factorization  A=U*D*U**H  or
                 A=L*D*L**H computed by ZHETRF and converted by ZSYCONV

       zhetrs_rook - compute the solution to  a  system  of  linear  equations
                 A*X=B  for  Hermitian  matrices  using factorization obtained
                 with one of the bounded  diagonal  pivoting  methods  (max  2
                 interchanges)

       zhfrk - perform a Hermitian rank-k operation for matrix in RFP format

       zhgeqz  - implement a single-shift version of the QZ method for finding
                 the  generalized  eigenvalues  w(i)=ALPHA(i)/BETA(i)  of  the
                 equation    det(  A-w(i)  B  ) = 0  If JOB='S', then the pair
                 (A,B) is simultaneously reduced to Schur form (i.e., A and  B
                 are both upper triangular) by applying one unitary tranforma-
                 tion (usually called Q) on  the  left  and  another  (usually
                 called Z) on the right

       zhpcon  -  estimate the reciprocal of the condition number of a complex
                 Hermitian  packed  matrix  A  using  the  factorization  A  =
                 U*D*U**H or A = L*D*L**H computed by ZHPTRF

       zhpev  - compute all the eigenvalues and, optionally, eigenvectors of a
                 complex Hermitian matrix in packed storage

       zhpevd - compute all the eigenvalues and, optionally, eigenvectors of a
                 complex Hermitian matrix A in packed storage

       zhpevx  - compute selected eigenvalues and, optionally, eigenvectors of
                 a complex Hermitian matrix A in packed storage

       zhpgst - reduce a complex Hermitian-definite  generalized  eigenproblem
                 to standard form, using packed storage

       zhpgv  -  compute all the eigenvalues and, optionally, the eigenvectors
                 of a complex generalized Hermitian-definite eigenproblem,  of
                 the     form     A*x=(lambda)*B*x,     A*Bx=(lambda)*x,    or
                 B*A*x=(lambda)*x

       zhpgvd - compute all the eigenvalues and, optionally, the  eigenvectors
                 of  a complex generalized Hermitian-definite eigenproblem, of
                 the    form     A*x=(lambda)*B*x,     A*Bx=(lambda)*x,     or
                 B*A*x=(lambda)*x

       zhpgvx  - compute selected eigenvalues and, optionally, eigenvectors of
                 a complex generalized Hermitian-definite eigenproblem, of the
                 form A*x=(lambda)*B*x, A*Bx=(lambda)*x, or B*A*x=(lambda)*x

       zhpmv - perform the matrix-vector operationy := alpha*A*x + beta*y

       zhpr  - perform the hermitian rank 1 operation   A := alpha*x*conjg( x'
                 ) + A

       zhpr2 - perform the Hermitian rank 2 operation   A := alpha*x*conjg( y'
                 ) + conjg( alpha )*y*conjg( x' ) + A

       zhprfs  - improve the computed solution to a system of linear equations
                 when the  coefficient  matrix  is  Hermitian  indefinite  and
                 packed, and provide error bounds and backward error estimates
                 for the solution

       zhpsv - compute the solution to a complex system of linear equationsA *
                 X = B, where A is an N-by-N Hermitian matrix stored in packed
                 format and X and B are N-by-NRHS matrices

       zhpsvx - use the diagonal pivoting factorization A = U*D*U**H  or  A  =
                 L*D*L**H  to compute the solution to a complex system of lin-
                 ear equations A * X = B,  where  A  is  an  N-by-N  Hermitian
                 matrix  stored  in  packed  format  and X and B are N-by-NRHS
                 matrices

       zhptrd - reduce a complex Hermitian matrix A stored in packed  form  to
                 real  symmetric  tridiagonal  form  T by a unitary similarity
                 transformation

       zhptrf - compute the factorization of a complex Hermitian packed matrix
                 A using the Bunch-Kaufman diagonal pivoting method

       zhptri - compute the inverse of a complex Hermitian indefinite matrix A
                 in packed storage using the factorization A = U*D*U**H or A =
                 L*D*L**H computed by ZHPTRF

       zhptrs - solve a system of linear equations A*X = B with a complex Her-
                 mitian matrix A stored in packed format using the  factoriza-
                 tion A = U*D*U**H or A = L*D*L**H computed by ZHPTRF

       zhsein  -  use  inverse  iteration  to find specified right and/or left
                 eigenvectors of a complex upper Hessenberg matrix H

       zhseqr - compute the eigenvalues of a complex upper  Hessenberg  matrix
                 H,  and,  optionally,  the  matrices  T  and Z from the Schur
                 decomposition H = Z T Z**H, where T is  an  upper  triangular
                 matrix (the Schur form), and Z is the unitary matrix of Schur
                 vectors

       zinfinite_norm_error - A utility function of the  SuperLU  solver  that
                 computes  the  infinity-norm  of an array of vectors that are
                 approximations to the exact solution vector.

       zjadmm - Jagged diagonal matrix-matrix multiply (modified Ellpack)

       zjadrp - right permutation of a jagged diagonal matrix

       zjadsm - Jagged-diagonal format triangular solve

       zla_gbamv - perform a matrix-vector operation to calculate error bounds

       zla_gbrcond_c   -   compute  the  infinity  norm  condition  number  of
                 op(A)*inv(diag(c)) for general banded matrices

       zla_gbrcond_x  -  compute  the  infinity  norm  condition   number   of
                 op(A)*diag(x) for general banded matrices

       zla_gbrfsx_extended - improve the computed solution to a system of lin-
                 ear equations  for  general  banded  matrices  by  performing
                 extra-precise  iterative  refinement and provide error bounds
                 and backward error estimates for the solution

       zla_gbrpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U) for a general banded matrix

       zla_geamv  -  compute a matrix-vector product using a general matrix to
                 calculate error bounds

       zla_gercond_c  -  compute  the  infinity  norm  condition   number   of
                 op(A)*inv(diag(c)) for general matrices

       zla_gercond_x   -   compute  the  infinity  norm  condition  number  of
                 op(A)*diag(x) for general matrices

       zla_gerfsx_extended - improve the computed solution to a system of lin-
                 ear  equations  by performing extra-precise iterative refine-
                 ment and provide error bounds and  backward  error  estimates
                 for the solution

       zla_gerpvgrw  -  compute  the  reciprocal pivot growth factor using the
                 "max absolute element" norm

       zla_heamv - compute a matrix-vector product using a  Hermitian  indefi-
                 nite matrix to calculate error bounds

       zla_hercond_c   -   compute  the  infinity  norm  condition  number  of
                 op(A)*inv(diag(c)) for Hermitian indefinite matrices

       zla_hercond_x  -  compute  the  infinity  norm  condition   number   of
                 op(A)*diag(x) for Hermitian indefinite matrices

       zla_herfsx_extended - improve the computed solution to a system of lin-
                 ear equations for Hermitian indefinite matrices by performing
                 extra-precise  iterative  refinement and provide error bounds
                 and backward error estimates for the solution

       zla_herpvgrw - compute the reciprocal pivot  growth  factor  using  the
                 "max absolute element" norm

       zla_lin_berr - compute a component-wise relative backward error

       zla_porcond_c   -   compute  the  infinity  norm  condition  number  of
                 op(A)*inv(diag(c)) for Hermitian positive-definite matrices

       zla_porcond_x  -  compute  the  infinity  norm  condition   number   of
                 op(A)*diag(x) for Hermitian positive-definite matrices

       zla_porfsx_extended - improve the computed solution to a system of lin-
                 ear equations for symmetric  or  Hermitian  positive-definite
                 matrices by performing extra-precise iterative refinement and
                 provide error bounds and backward  error  estimates  for  the
                 solution

       zla_porpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U) for a symmetric or  Hermitian  positive-defi-
                 nite matrix

       zla_syamv  -  compute a matrix-vector product using a symmetric indefi-
                 nite matrix to calculate error bounds

       zla_syrcond_c  -  compute  the  infinity  norm  condition   number   of
                 op(A)*inv(diag(c)) for symmetric indefinite matrices

       zla_syrcond_x   -   compute  the  infinity  norm  condition  number  of
                 op(A)*diag(x) for symmetric indefinite matrices

       zla_syrfsx_extended - improve the computed solution to a system of lin-
                 ear equations for symmetric indefinite matrices by performing
                 extra-precise iterative refinement and provide  error  bounds
                 and backward error estimates for the solution

       zla_syrpvgrw   -   compute   the   reciprocal   pivot   growth   factor
                 norm(A)/norm(U) for a symmetric indefinite matrix

       zla_wwaddw - add a vector into a doubled-single vector

       zlaed0 - is used by sstedc. Compute all eigenvalues  and  corresponding
                 eigenvectors  of  an  unreduced  symmetric tridiagonal matrix
                 using the divide and conquer method

       zlaed7 - is used by sstedc. Compute the updated eigensystem of a diago-
                 nal matrix after modification by a rank-one symmetric matrix.
                 Used when the original matrix is dense

       zlaed8 - is used by sstedc.  Merge  eigenvalues  and  deflates  secular
                 equation. Used when the original matrix is dense

       zlahef - compute a partial factorization of a complex Hermitian indefi-
                 nite matrix using the Bunch-Kaufman diagonal pivoting  method
                 (blocked algorithm, calling Level 3 BLAS)

       zlahef_rook  -  compute  a partial factorization of a complex Hermitian
                 indefinite matrix using the  bounded  Bunch-Kaufman  ("rook")
                 diagonal  pivoting method (blocked algorithm, calling Level 3
                 BLAS)

       zlals0 - apply back multiplying factors in solving  the  least  squares
                 problem using divide and conquer SVD approach. Used by sgelsd

       zlalsa - compute the SVD of the coefficient  matrix  in  compact  form.
                 Used by sgelsd

       zlalsd  -  use the singular value decomposition of A to solve the least
                 squares problem

       zlangs - returns the value of the one-norm, or the  Frobenius-norm,  or
                 the infinity-norm, or the element with largest absolute value
                 of a general real matrix A in sparse format.

       zlanhf - return the value of the 1-norm, or the Frobenius norm, or  the
                 infinity  norm, or the element of largest absolute value of a
                 Hermitian matrix in RFP format

       zlaqgs - a SuperLU function that equilibrates a general sparse M  by  N
                 matrix A.

       zlarscl2 - perform reciprocal diagonal scaling on a vector

       zlarz  -  apply  a  complex  elementary reflector H to a complex M-by-N
                 matrix C, from either the left or the right

       zlarzb - apply a complex block reflector H or its transpose H**H  to  a
                 complex distributed M-by-N C from the left or the right

       zlarzt - form the triangular factor T of a complex block reflector H of
                 order > n, which is defined as  a  product  of  k  elementary
                 reflectors

       zlascl2 - perform diagonal scaling on a vector

       zlasyf  - compute a partial factorization of a complex symmetric matrix
                 using the Bunch-Kaufman diagonal pivoting method

       zlasyf_rook - compute a partial factorization of  a  complex  symmetric
                 matrix using the bounded Bunch-Kaufman ("rook") diagonal piv-
                 oting method

       zlatrz - factor an upper trapezoidal matrix by means of unitary  trans-
                 formations

       zlatzm - routine is deprecated and has been replaced by routine ZUNMRZ

       zpbcon  -  estimate  the  reciprocal  of  the  condition number (in the
                 1-norm) of a complex Hermitian positive definite band  matrix
                 using  the  Cholesky  factorization  A = U**H*U or A = L*L**H
                 computed by ZPBTRF

       zpbequ - compute row and column scalings intended to equilibrate a Her-
                 mitian  positive definite band matrix A and reduce its condi-
                 tion number (with respect to the two-norm)

       zpbrfs - improve the computed solution to a system of linear  equations
                 when  the  coefficient  matrix is Hermitian positive definite
                 and banded, provide error bounds and backward error estimates
                 for the solution

       zpbstf  - compute a split Cholesky factorization of a complex Hermitian
                 positive definite band matrix A

       zpbsv - compute the solution to a complex system  of  linear  equations
                 A*X=B,  where A is an N-by-N Hermitian positive definite band
                 matrix and X and B are N-by-NRHS matrices

       zpbsvx - use the Cholesky factorization A=U**H*U or A=L*L**H to compute
                 the  solution  to a complex system of linear equations A*X=B,
                 where A is an N-by-N Hermitian positive definite band  matrix
                 and X and B are N-by-NRHS matrices

       zpbtf2 - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite band matrix A

       zpbtrf - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite band matrix A

       zpbtrs  -  solve  a system of linear equations A*X = B with a Hermitian
                 positive definite band matrix A using the Cholesky factoriza-
                 tion A = U**H*U or A = L*L**H computed by ZPBTRF

       zpftrf  -  computes  the  Cholesky factorization of a complex Hermitian
                 positive definite matrix A, the block version  of  the  algo-
                 rithm

       zpftri  -  compute the inverse of a complex Hermitian positive definite
                 matrix A using the Cholesky factorization computed by ZPFTRF

       zpftrs - solve a system of linear equations A*X = B  with  a  Hermitian
                 positive  definite  matrix A using the Cholesky factorization
                 computed by ZPFTRF

       zpocon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm) of a complex Hermitian positive definite matrix using
                 the Cholesky factorization A = U**H*U or A = L*L**H  computed
                 by ZPOTRF

       zpoequ - compute row and column scalings intended to equilibrate a Her-
                 mitian positive definite matrix A and  reduce  its  condition
                 number (with respect to the two-norm)

       zpoequb  -  compute  row  and column scalings intended to equilibrate a
                 symmetric positive definite matrix A and reduce its condition
                 number with respect to the two-norm

       zporfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is Hermitian  positive  definite,
                 provide  error  bounds  and  backward error estimates for the
                 solution

       zporfsx - improve the computed solution to a system of linear equations
                 when  the  coefficient matrix is symmetric positive definite,
                 provide error bounds and backward  error  estimates  for  the
                 solution

       zposv - compute the solution to a complex system of linear equationsA*X
                 = B, where A is an N-by-N Hermitian positive definite  matrix
                 and X and B are N-by-NRHS matrices

       zposvx  -  use  the Cholesky factorization to compute the solution to a
                 complex system of linear equations  A*X = B, where A is an N-
                 by-N Hermitian positive definite matrix and X and B are N-by-
                 NRHS matri ces

       zposvxx - compute the solution to a complex system of linear  equations
                 A*X  =  B,  where  A is an N-by-N symmetric positive definite
                 matrix and X and B are N-by-NRHS matrices

       zpotf2 - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite matrix A

       zpotrf - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite matrix A

       zpotri - compute the inverse of a complex Hermitian  positive  definite
                 matrix  A  using the Cholesky factorization A = U**H*U or A =
                 L*L**H computed by ZPOTRF

       zpotrs - solve a system of linear equations A*X = B  with  a  Hermitian
                 positive definite matrix A using the Cholesky factorization A
                 = U**H*U or A = L*L**H computed by ZPOTRF

       zppcon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm)  of  a  complex  Hermitian  positive  definite packed
                 matrix using the Cholesky factorization A =  U**H*U  or  A  =
                 L*L**H computed by ZPPTRF

       zppequ - compute row and column scalings intended to equilibrate a Her-
                 mitian positive definite  matrix  A  in  packed  storage  and
                 reduce its condition number (with respect to the two-norm)

       zpprfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is  Hermitian  positive  definite
                 and packed, provide error bounds and backward error estimates
                 for the solution

       zppsv - compute the solution to a complex system  of  linear  equations
                 A*X  =  B,  where  A is an N-by-N Hermitian positive definite
                 matrix stored in packed format and  X  and  B  are  N-by-NRHS
                 matrices

       zppsvx  -  use  the  Cholesky factorization A = U**H*U or A = L*L**H to
                 compute the solution to a complex system of linear  equations
                 A * X = B,

       zpptrf - compute the Cholesky factorization of a complex Hermitian pos-
                 itive definite matrix A stored in packed format

       zpptri - compute the inverse of a complex Hermitian  positive  definite
                 matrix  A  using the Cholesky factorization A = U**H*U or A =
                 L*L**H computed by ZPPTRF

       zpptrs - solve a system of linear equations A*X = B  with  a  Hermitian
                 positive  definite  matrix  A  in  packed  storage  using the
                 Cholesky factorization A = U**H*U or A = L*L**H  computed  by
                 ZPPTRF

       zpstf2 - compute the Cholesky factorization with complete pivoting of a
                 complex Hermitian positive semidefinite matrix A

       zpstrf - compute the Cholesky factorization with complete pivoting of a
                 complex Hermitian positive semidefinite matrix A

       zptcon - compute the reciprocal of the condition number (in the 1-norm)
                 of a complex Hermitian positive definite  tridiagonal  matrix
                 using the factorization A = L*D*L**H or A = U**H*D*U computed
                 by ZPTTRF

       zpteqr - compute all eigenvalues and,  optionally,  eigenvectors  of  a
                 symmetric  positive definite tridiagonal matrix by first fac-
                 toring the matrix using SPTTRF and  then  calling  CBDSQR  to
                 compute the singular values of the bidiagonal factor

       zptrfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is  Hermitian  positive  definite
                 and  tridiagonal,  provide  error  bounds  and backward error
                 estimates for the solution

       zptsv - compute the solution to a complex system  of  linear  equations
                 A*X  =  B,  where  A is an N-by-N Hermitian positive definite
                 tridiagonal matrix, and X and B are N-by-NRHS matrices

       zptsvx - use the factorization A = L*D*L**H to compute the solution  to
                 a  complex  system of linear equations A*X = B, where A is an
                 N-by-N Hermitian positive definite tridiagonal matrix  and  X
                 and B are N-by-NRHS matrices

       zpttrf  - compute the L*D*L' factorization of a complex Hermitian posi-
                 tive definite tridiagonal matrix A

       zpttrs - solve a tridiagonal system of the form  A * X =  B  using  the
                 factorization A = U'*D*U or A = L*D*L' computed by ZPTTRF

       zptts2  -  solve  a tridiagonal system of the form  A * X = B using the
                 factorization A = U'*D*U or A = L*D*L' computed by ZPTTRF

       zrot - apply a plane rotation, where the cos (C) is real  and  the  sin
                 (S) is complex, and the vectors X and Y are complex

       zrotg - Construct a Given's plane rotation

       zscal - Compute y := alpha * y

       zsctr - Scatters elements from x into y

       zskymm - Skyline format matrix-matrix multiply

       zskysm - Skyline format triangular solve

       zspcon  -  estimate  the  reciprocal  of  the  condition number (in the
                 1-norm) of a complex symmetric packed matrix A using the fac-
                 torization A = U*D*U**T or A = L*D*L**T computed by ZSPTRF

       zsprfs  - improve the computed solution to a system of linear equations
                 when the  coefficient  matrix  is  symmetric  indefinite  and
                 packed, provide error bounds and backward error estimates for
                 the solution

       zspsv - compute the solution to a complex system of linear equationsA*X
                 =  B,  where A is an N-by-N symmetric matrix stored in packed
                 format and X and B are N-by-NRHS matrices

       zspsvx - use the diagonal pivoting factorization A = U*D*U**T  or  A  =
                 L*D*L**T  to compute the solution to a complex system of lin-
                 ear equations A * X = B,  where  A  is  an  N-by-N  symmetric
                 matrix  stored  in  packed  format  and X and B are N-by-NRHS
                 matrices

       zsptrf - compute the factorization of  a  complex  symmetric  matrix  A
                 stored in packed format using the Bunch-Kaufman diagonal piv-
                 oting method

       zsptri - compute the inverse of a complex symmetric indefinite matrix A
                 in packed storage using the factorization A = U*D*U**T or A =
                 L*D*L**T computed by ZSPTRF

       zsptrs - solve a system of linear equations A*X = B with a complex sym-
                 metric  matrix A stored in packed format using the factoriza-
                 tion A = U*D*U**T or A = L*D*L**T computed by ZSPTRF

       zstedc - compute all eigenvalues and,  optionally,  eigenvectors  of  a
                 symmetric  tridiagonal  matrix  using  the divide and conquer
                 method

       zstegr - Compute T-sigma_i = L_i D_i L_i^T, such that L_i D_i L_i^T  is
                 a relatively robust representation

       zstein  -  compute  the  eigenvectors  of  a real symmetric tridiagonal
                 matrix  T  corresponding  to  specified  eigenvalues,   using
                 inverse iteration

       zstemr -  compute selected eigenvalues and, optionally, eigenvectors of
                 a real symmetric tridiagonal matrix T

       zsteqr - compute all eigenvalues and,  optionally,  eigenvectors  of  a
                 symmetric  tridiagonal  matrix  using  the  implicit QL or QR
                 method

       zstsv - compute the solution to a complex system of linear equations  A
                 * X = B where A is a symmetric tridiagonal matrix

       zsttrf  -  compute the factorization of a complex symmetric tridiagonal
                 matrix A using the Bunch-Kaufman diagonal pivoting method

       zsttrs - compute the solution to a complex system of  linear  equations
                 A*X  =  B,  where A is an N-by-N symmetric tridiagonal matrix
                 and X and B are N-by-NRHS matrices

       zswap - Exchange vectors x and y

       zsycon - estimate the  reciprocal  of  the  condition  number  (in  the
                 1-norm)  of a complex symmetric matrix A using the factoriza-
                 tion A = U*D*U**T or A = L*D*L**T computed by ZSYTRF

       zsycon_rook - estimate the reciprocal of the condition number  (in  the
                 1-norm) of a complex symmetric matrix using the factorization
                 A = U*D*U**T or A = L*D*L**T computed by ZSYTRF_ROOK

       zsyconv - converts A given by ZHETRF into L and D or vice-versa

       zsyequb - compute row and column scalings  intended  to  equilibrate  a
                 symmetric  matrix  A  and  reduce  its  condition number with
                 respect to the two-norm

       zsymm - perform one of the matrix-matrix  operationsC  :=  alpha*A*B  +
                 beta*C or C := alpha*B*A + beta*C

       zsyr2k  -  perform  one  of  the  symmetric  rank  2k operations   C :=
                 alpha*A*B' +  alpha*B*A'  +  beta*C  or  C  :=  alpha*A'*B  +
                 alpha*B'*A + beta*C

       zsyrfs  - improve the computed solution to a system of linear equations
                 when the coefficient matrix is symmetric indefinite,  provide
                 error bounds and backward error estimates for the solution

       zsyrfsx - improve the computed solution to a system of linear equations
                 when the coefficient matrix is symmetric indefinite, provides
                 error bounds and backward error estimates for the solution

       zsyrk  -  perform  one  of  the  symmetric  rank  k  operations    C :=
                 alpha*A*A' + beta*C or C := alpha*A'*A + beta*C

       zsysv - compute the solution to a complex system of linear equationsA*X
                 = B, where A is an N-by-N symmetric matrix and X and B are N-
                 by-NRHS matrices

       zsysv_rook - compute the solution to system of linear equations A*X = B
                 for  symmetric matrices. ZSYTRF_ROOK is called to compute the
                 factorization of A

       zsysvx - use the diagonal pivoting factorization to compute  the  solu-
                 tion to a complex system of linear equations A*X = B, where A
                 is an N-by-N symmetric matrix  and  X  and  B  are  N-by-NRHS
                 matrices

       zsysvxx  -  compute  the solution to complex system of linear equations
                 A*X = B for symmetric matrices

       zsytf2 - compute the factorization of  a  complex  symmetric  matrix  A
                 using the Bunch-Kaufman diagonal pivoting method

       zsytf2_rook  - compute the factorization of a complex symmetric indefi-
                 nite matrix using the bounded Bunch-Kaufman ("rook") diagonal
                 pivoting method (unblocked algorithm)

       zsytrf  -  compute  the  factorization  of a complex symmetric matrix A
                 using the Bunch-Kaufman diagonal pivoting method

       zsytrf_rook - compute the factorization of a complex  symmetric  matrix
                 using  the  bounded  Bunch-Kaufman ("rook") diagonal pivoting
                 method (blocked algorithm)

       zsytri - compute the inverse of a complex symmetric indefinite matrix A
                 using the factorization A = U*D*U**T or A = L*D*L**T computed
                 by ZSYTRF

       zsytri2 - compute the inverse  of  a  COMPLEX*16  symmetric  indefinite
                 matrix A using the factorization A = U*D*U**T or A = L*D*L**T
                 computed by ZSYTRF

       zsytri2x - compute the inverse of a complex symmetric indefinite matrix
                 A using the factorization computed by ZSYTRF

       zsytri_rook  -  compute  the  inverse of a complex symmetric indefinite
                 matrix A using the factorization A = U*D*U**T or A = L*D*L**T
                 computed by ZSYTRF_ROOK

       zsytrs - solve a system of linear equations A*X = B with a complex sym-
                 metric matrix A using the factorization A = U*D*U**T or  A  =
                 L*D*L**T computed by ZSYTRF

       zsytrs2  -  solve  a  system of linear equations A*X = B with a complex
                 symmetric matrix A using the factorization computed by ZSYTRF
                 and converted by ZSYCONV

       zsytrs_rook - solve a system of linear equations A*X = B with a complex
                 symmetric matrix A using the factorization A = U*D*U**T or  A
                 = L*D*L**T computed by ZSYTRF_ROOK

       ztbcon  - estimate the reciprocal of the condition number of a triangu-
                 lar band matrix A, in either the 1-norm or the infinity-norm

       ztbmv - perform one of the matrix-vector operationsx := A*x,  or  x  :=
                 A'*x, or x := conjg( A' )*x

       ztbrfs  -  provide  error  bounds  and backward error estimates for the
                 solution to a system of linear equations  with  a  triangular
                 band coefficient matrix

       ztbsv  -  solve one of the systems of equations   A*x = b, or A'*x = b,
                 or conjg( A' )*x = b

       ztbtrs - solve a triangular system of the form A*X = B, A**T*X = B,  or
                 A**H*X  =  B, where A is a triangular band matrix of order N,
                 and B is an N-by-NRHS matrix

       ztfsm - solve a matrix equation (one operand is a triangular matrix  in
                 RFP format)

       ztftri  -  compute  the  inverse of a triangular matrix A stored in RFP
                 format

       ztfttp - copy a triangular matrix from the rectangular full packed for-
                 mat (TF) to the standard packed format (TP)

       ztfttr - copy a triangular matrix from the rectangular full packed for-
                 mat (TF) to the standard full format (TR)

       ztgevc - compute the right and/or left generalized  eigenvectors  of  a
                 pair  of complex upper triangular matrices with real diagonal
                 elements (A,B) obtained from the generalized Schur factoriza-
                 tion  of  an  original  pair of complex nonsymmetric matrices
                 (AO,BO)

       ztgexc - reorder the  generalized  Schur  decomposition  of  a  complex
                 matrix  pair  (A,B), using an unitary equivalence transforma-
                 tion (A, B) := Q * (A, B) * Z', so that the diagonal block of
                 (A, B) with row index IFST is moved to row ILST

       ztgsen  -  reorder  the  generalized  Schur  decomposition of a complex
                 matrix pair (A, B), so that a selected cluster of eigenvalues
                 appears in the leading diagonal blocks of the pair (A,B)

       ztgsja - compute the generalized singular value decomposition (GSVD) of
                 two complex upper triangular (or trapezoidal) matrices A  and
                 B

       ztgsna  - estimate reciprocal condition numbers for specified eigenval-
                 ues and/or eigenvectors of a matrix pair (A, B)

       ztgsyl - solve the generalized Sylvester equation

       ztpcon - estimate the reciprocal of the condition number  of  a  packed
                 triangular  matrix  A,  in either the 1-norm or the infinity-
                 norm

       ztpmqrt - apply a complex orthogonal matrix Q obtained from a "triangu-
                 lar-pentagonal"  complex  block reflector H to a general com-
                 plex matrix C, which consists of two blocks

       ztpmv - perform one of the matrix-vector operations x := A*x, or  x  :=
                 A'*x, or x := conjg( A' )*x

       ztpqrt  -  compute a blocked QR factorization of a complex "triangular-
                 pentagonal" matrix C, which is composed of a triangular block
                 A and pentagonal block B, using the compact WY representation
                 for Q

       ztpqrt2 - compute a QR factorization of a real or complex  "triangular-
                 pentagonal"  matrix,  which is composed of a triangular block
                 and a pentagonal block, using the compact  WY  representation
                 for Q

       ztprfs  -  provide  error  bounds  and backward error estimates for the
                 solution to a system of linear equations  with  a  triangular
                 packed coefficient matrix

       ztpsv - solve one of the systems of equations  A*x = b, or A'*x = b, or
                 conjg( A' )*x = b

       ztptri - compute the inverse of a complex  upper  or  lower  triangular
                 matrix A stored in packed format

       ztptrs  - solve a triangular system of the form A*X = B, A**T*X = B, or
                 A**H*X = B, where A is a triangular matrix of order N  stored
                 in packed format, and B is an N-by-NRHS matrix

       ztpttf  - copy a triangular matrix from the standard packed format (TP)
                 to the rectangular full packed format (TF)

       ztpttr - copy a triangular matrix from the standard packed format  (TP)
                 to the standard full format (TR)

       ztrans - transpose and scale source matrix

       ztrcon  - estimate the reciprocal of the condition number of a triangu-
                 lar matrix A, in either the 1-norm or the infinity-norm

       ztrevc - compute some or all of the right and/or left eigenvectors of a
                 complex upper triangular matrix T

       ztrexc  -  reorder  the  Schur  factorization  of  a complex matrix A =
                 Q*T*Q**H, so that the diagonal element of T  with  row  index
                 IFST is moved to row ILST

       ztrmm  -  perform  one of the matrix-matrix operations B := alpha*op( A
                 )*B, or B := alpha*B*op( A ) where alpha is a scalar, B is an
                 m  by n matrix, A is a unit, or non-unit, upper or lower tri-
                 angular matrix and op( A ) is one of op( A ) = A or op( A ) =
                 A' or op( A ) = conjg( A' )

       ztrmv  -  perform one of the matrix-vector operations x := A*x, or x :=
                 A'*x, or x := conjg( A' )*x

       ztrrfs - provide error bounds and  backward  error  estimates  for  the
                 solution  to  a  system of linear equations with a triangular
                 coefficient matrix

       ztrsen - reorder the Schur  factorization  of  a  complex  matrix  A  =
                 Q*T*Q**H,  so  that a selected cluster of eigenvalues appears
                 in the leading positions on the diagonal of the upper  trian-
                 gular  matrix  T,  and  the  leading  columns  of  Q  form an
                 orthonormal basis of the corresponding right  invariant  sub-
                 space

       ztrsm - solve one of the matrix equations op( A )*X = alpha*B, or X*op(
                 A ) = alpha*B

       ztrsna - estimate reciprocal condition numbers for specified  eigenval-
                 ues  and/or  right eigenvectors of a complex upper triangular
                 matrix T (or of any matrix Q*T*Q**H with Q unitary)

       ztrsv - solve one of the systems of equations  A*x = b, or A'*x = b, or
                 conjg(A')*x = b

       ztrsyl - solve the complex Sylvester matrix equation

       ztrti2  -  compute  the  inverse of a complex upper or lower triangular
                 matrix

       ztrtri - compute the inverse of a complex  upper  or  lower  triangular
                 matrix A

       ztrtrs  - solve a triangular system of the form A*X = B, A**T*X = B, or
                 A**H*X = B, where A is a triangular matrix of order N, and  B
                 is an N-by-NRHS matrix

       ztrttf - copy a triangular matrix from the standard full format (TR) to
                 the rectangular full packed format (TF)

       ztrttp - copy a triangular matrix from the standard full format (TR) to
                 the standard packed format (TP)

       ztzrqf - routine is deprecated and has been replaced by routine ZTZRZF

       ztzrzf  - reduce the M-by-N ( M<=N ) complex upper trapezoidal matrix A
                 to upper triangular form by means of unitary transformations

       zunbdb - simultaneously bidiagonalize the blocks of  an  M-by-M  parti-
                 tioned unitary matrix

       zunbdb1  - simultaneously bidiagonalize the blocks of a tall and skinny
                 matrix with orthonomal columns

       zunbdb2 - simultaneously bidiagonalize the blocks of a tall and  skinny
                 matrix with orthonomal columns

       zunbdb3  - simultaneously bidiagonalize the blocks of a tall and skinny
                 matrix with orthonomal columns

       zunbdb4 - simultaneously bidiagonalize the blocks of a tall and  skinny
                 matrix with orthonomal columns

       zunbdb5 - orthogonalize one column vector with respect to the orthonor-
                 mal columns of another column vector

       zunbdb6 - orthogonalize one column vector with respect to the orthonor-
                 mal columns of another column vector

       zuncsd  - compute the CS decomposition of an M-by-M partitioned unitary
                 matrix

       zuncsd2by1 - compute the CS decomposition  of  an  M-by-Q  matrix  with
                 orthonormal  columns  that has been partitioned into a 2-by-1
                 block structure

       zung2l - generate all or part of the unitary matrix Q from a QL factor-
                 ization determined by cgeqlf (unblocked algorithm)

       zung2r  - generate an M-by-N complex matrix Q with orthonormal columns,
                 which is defined as the first N columns of  a  product  of  N
                 elementary reflectors of order M

       zungbr  - generate one of the complex unitary matrices Q or P**H deter-
                 mined by ZGEBRD when reducing a complex matrix A to  bidiago-
                 nal form

       zunghr  -  generate  a complex unitary matrix Q which is defined as the
                 product of IHI-ILO  elementary  reflectors  of  order  N,  as
                 returned by ZGEHRD

       zungl2  -  generate all or part of the unitary matrix Q from an LQ fac-
                 torization determined by zgelqf (unblocked algorithm)

       zunglq - generate an M-by-N complex matrix  Q  with  orthonormal  rows,
                 which  is  defined as the first M rows of a product of K ele-
                 mentary reflectors of order N

       zungql - generate an M-by-N complex matrix Q with orthonormal  columns,
                 which is defined as the last N columns of a product of K ele-
                 mentary reflectors of order M

       zungqr - generate an M-by-N complex matrix Q with orthonormal  columns,
                 which  is  defined  as  the first N columns of a product of K
                 elementary reflectors of order M

       zungr2 - generate all or part of the unitary matrix Q from an  RQ  fac-
                 torization determined by cgerqf (unblocked algorithm)

       zungrq  -  generate  an  M-by-N complex matrix Q with orthonormal rows,
                 which is defined as the last M rows of a product of K elemen-
                 tary reflectors of order N

       zungtr  -  generate  a complex unitary matrix Q which is defined as the
                 product of n-1 elementary reflectors of order N, as  returned
                 by CHETRD

       zunm2l - multiply a general matrix by the unitary matrix from a QL fac-
                 torization determined by cgeqlf (unblocked algorithm)

       zunm2r - multipliy a general matrix by the unitary  matrix  from  a  QR
                 factorization determined by cgeqrf (unblocked algorithm)

       zunmbr - overwrite the general complex M-by-N matrix C Q*C or Q**H*C or
                 C*Q**H or C*Q or P*C or P**H*C or C*P or C*P**H

       zunmhr - overwrite the general complex M-by-N  matrix  C  with  Q*C  or
                 Q**H*C or C*Q**H or C*Q

       zunml2 - multiply a general matrix by the unitary matrix from a LQ fac-
                 torization determined by cgelqf (unblocked algorithm)

       zunmlq - overwrite the general complex M-by-N matrix  C  with  Q*C,  or
                 Q**H*C,  or  C*Q**H,  or  C*Q,  where  Q is a complex unitary
                 matrix defined as the product of K elementary reflectors

       zunmql - overwrite the general complex M-by-N matrix  C  with  Q*C,  or
                 Q**H*C,  or  C*Q**H,  or  C*Q,  where  Q is a complex unitary
                 matrix defined as the product of K elementary reflectors

       zunmqr - overwrite the general complex M-by-N matrix C  with    SIDE  =
                 'L' SIDE = 'R' TRANS = 'N'

       zunmr2 - multiply a general matrix by the unitary matrix from a RQ fac-
                 torization determined by zgerqf (unblocked algorithm)

       zunmr3 - multiply a general matrix by the unitary matrix from a RZ fac-
                 torization determined by ctzrzf (unblocked algorithm)

       zunmrq  -  overwrite  the  general complex M-by-N matrix C with Q*C, or
                 Q**H*C, or C*Q**H, or C*Q,  where  Q  is  a  complex  unitary
                 matrix defined as the product of K elementary reflectors

       zunmrz  -  overwrite  the  general  complex M-by-N matrix C with Q*C or
                 Q**H*C or C*Q**H or C*Q.

       zunmtr - overwrite the general complex M-by-N matrix  C  with  Q*C,  or
                 Q**H*C,  or C*Q**H, or C*Q, where Q is defined as the product
                 of elementary reflectors, as returned by ZHETRD

       zupgtr - generate a complex unitary matrix Q which is  defined  as  the
                 product  of  n-1  elementary  reflectors  H(i) of order n, as
                 returned by CHPTRD using packed storage

       zupmtr - overwrite the general complex M-by-N matrix C  with    SIDE  =
                 'L' SIDE = 'R' TRANS = 'N'

       zvbrmm - variable block sparse row format matrix-matrix multiply

       zvbrsm - variable block sparse row format triangular solve

       zvmul - compute the scaled product of complex vectors



                                  7 Nov 2015                         intro(3P)