Numerical Computation Guide

A - B - C - D - E - F - G - H - I - J - K - L - M - N - O - P - Q - R - S - T - U - V - W - X - Y - Z

# Index

### A

abort on exception
C example, 1
accuracy, 1
floating-point operations, 1
significant digits (number of), 1
threshold, 1
`adb`, 1
`addrans`
random number utilities, 1
argument reduction
trigonometric functions, 1

### B

base conversion
base 10 to base 2, 1
base 2 to base 10, 1
formatted I/O, 1
Bessel functions, 1

### C

C driver
example, call FORTRAN subroutines from C, 1
clock speed, 1
conversion between number sets, 1
conversions between decimal strings and binary floating-point numbers, 1
`convert_external`
binary floating-point, 1
data conversion, 1

### D

data types
relation to IEEE formats, 1
`dbx`, 1
decimal representation
maximum positive normal number, 1
minimum positive normal number, 1
precision, 1
ranges, 1
double-precision representation
C example, 1
FORTRAN example, 1

### E

`errno.h`
define values for `errno`, 1
examine the accrued exception bits
C example, 1
examine the accrued exception flags
C example, 1

### F

`f77_floatingpoint.h`
define handler types
FORTRAN, 1
`-fast`, 1
floating-point
exceptions list, 1
rounding direction, 1
rounding precision, 1
tutorial, 1
floating-point accuracy
decimal strings and binary floating-point numbers, 1
floating-point exceptions, 1
abort on exceptions, 1
accrued exception bits, 1
common exceptions, 1
default result, 1
definition, 1
flags, 1
accrued, 1
current, 1
`ieee_functions`, 1
`ieee_retrospective`, 1
list of exceptions, 1
priority, 1
trap precedence, 1
floating-point options, 1
floating-point queue (FQ), 1
floating-point status register (FSR), 1, 2
`floatingpoint.h`
define handler types
C and C++, 1
flush to zero (see `Store 0`), 1
`fmod`, 1
`-fnonstd`, 1

### G

generate an array of numbers
FORTRAN example, 1
Goldberg paper, 1
abstract, 1
acknowledgments, 1
details, 1
IEEE standard, 1
IEEE standards, 1
introduction, 1
references, 1
rounding error, 1
summary, 1
systems aspects, 1
error properties, 1

### H

`HUGE`
compatibility with IEEE standard, 1
`HUGE_VAL`
compatibility with IEEE standard, 1

### I

IEEE double extended format
biased exponent
x86 architecture, 1
bit-field assignment
x86 architecture, 1
fraction
x86 architecture, 1
Inf
SPARC architecture, 1
x86 architecture, 1
NaN
x86 architecture, 1
normal number
SPARC architecture, 1
x86 architecture, 1
SPARC architecture, 1
sign bit
x86 architecture, 1
significand
explicit leading bit (x86 architecture), 1
subnormal number
SPARC architecture, 1
x86 architecture, 1
IEEE double format
biased exponent, 1
bit patterns and equivalent values, 1
bit-field assignment, 1
denormalized number, 1
fraction, 1
storage on SPARC, 1
storage on x86, 1
implicit bit, 1
Inf, infinity, 1
NaN, not a number, 1
normal number, 1
precision, 1
sign bit, 1
significand, 1
subnormal number, 1
IEEE formats
relation to language data types, 1
IEEE single format
biased exponent, 1
biased exponent,implicit bit, 1
bit assignments, 1
bit patterns and equivalent values, 1
bit-field assignment, 1
denormalized number, 1
fraction, 1
Inf,positive infinity, 1
mixed number, significand, 1
NaN, not a number, 1
normal number
maximum positive, 1
normal number bit pattern, 1
precision, normal number, 1
sign bit, 1
subnormal number bit pattern, 1
IEEE Standard 754
double extended format, 1
double format, 1
single format, 1
`ieee_flags`
accrued exception flag, 1
examine accrued exception bits-C example, 1
rounding direction, 1
rounding precision, 1, 2
set exception flags-C example, 1
truncate rounding, 1
`ieee_functions`
floating-point exceptions, 1
`ieee_handler`
abort on exception
FORTRAN example, 1
trap on exception
C example, 1
`ieee_handler`, 1
example, calling sequence, 1
trap on common exceptions, 1
`ieee_retrospective`
check underflow exception flag, 1
floating-point exceptions, 1
floating-point status register (FSR), 1
getting information about nonstandard IEEE modes, 1
getting information about outstanding exceptions, 1
`nonstandard_arithmetic` in effect, 1
precision, 1
rounding, 1
suppress exception messages, 1
`ieee_sun`
IEEE classification functions, 1
`ieee_values`
representing floating-point values, 1
representing Inf, 1
representing NaN, 1
representing normal number, 1
single-precision values, 1
`ieee_values` functions
C example, 1
Inf, 1, 2
default result of divide by zero, 1

### L

`lcrans`
random number utilities, 1
`libm`
default directories-executables, 1
list of functions, 1
standard installation, 1
SVID compliance, 1
`libm` functions
double precision, 1
single precision, 1
`libmil` (see also in-line templates), 1
`libsunmath`
default directories-executables, 1
list of functions, 1
standard installation, 1

### M

`MAXFLOAT`, 1

### N

`NaN`, 1, 2, 3, 4
`nonstandard_arithmetic`
turn off IEEE gradual underflow, 1
underflow, 1
normal number
maximum positive, 1
minimum positive, 1, 2
number line
binary representation, 1
decimal representation, 1
powers of 2, 1

### O

operating system math library
`libm.a`, 1
`libm.so`, 1

### P

pi
infinitely precise value, 1

### Q

quiet `NaN`
default result of invalid operation, 1

### R

random number generators, 1
random number utilities
`shufrans`, 1
represent double-precision value
C example, 1
FORTRAN example, 1
represent single-precision value
C example, 1
rounding direction, 1
C example, 1
rounding precision, 1
roundoff error
accuracy
loss of, 1

### S

set exception flags
C example, 1
`shufrans`
shuffle pseudo-random numbers, 1
single format, 1
single precision representation
C example, 1
Solaris versions supported, 1
square root instruction, 1, 2
`standard_arithmetic`
turn on IEEE behavior, 1
`Store 0`, 1
flush underflow results, 1, 2
subnormal number, 1
floating-point calculations, 1
SVID behavior of `libm`
`-Xt` compiler option, 1
SVID exceptions
`errno` set to `EDOM`
improper operands, 1
errno set to `ERANGE`
overflow or underflow, 1
`matherr`, 1
`PLOSS`, 1
`TLOSS`, 1
System V Interface Definition (SVID), 1

### T

trap
abort on exception, 1
`ieee_retrospective`, 1
trap on exception
C example, 1, 2
trap on floating-point exceptions
C example, 1
trigonometric functions
argument reduction, 1
tutorial, floating-point, 1

### U

underflow
floating-point calculations, 1
`nonstandard_arithmetic`, 1
threshold, 1
underflow thresholds
double extended precision, 1
double precision, 1
single precision, 1
unordered comparison
floating-point values, 1
`NaN`, 1

### V

`values.h`
define error messages, 1

### X

X/Open behavior of `libm`
`-Xa` compiler option, 1
`X_TLOSS`, 1
`-Xa`, 1
`-Xc`, 1
`-Xs`, 1
`-Xt`, 1