HP-UX Floating-Point Guide

240 Index
Index
operation errors. See invalid
operation conditions
operations
floating-point
, 61
moving out of loops
, 175
optimization
effect on application results
, 79
effect on performance
, 171
fenv(5) functions and
, 128
profile-based (PBO)
, 178
rounding mode and
, 56
options. See compiler options,
linker options
overflow
integer
, 165
overflow condition
, 56
results based on rounding
mode
, 57
P
PA-RISC architecture
determining type of
, 26
floating-point status register
,
128
math libraries
, 25, 112
Pascal math library
, 112
PBO (profile-based optimization)
and performance
, 178
performance analysis tools
, 169
performance tuning
, 167
position-independent code
and performance
, 178
positive infinity
round toward
, 55
positive zero
, 46
POSIX standard
, 24
precision
double. See double-precision
mixing, effect on performance
,
184
of floating-point values
, 40
quad. See quad-precision
single. See single-precision
prof(1)
profiling tool
, 169
profile-based optimization (PBO)
and performance
, 178
profiling tools
prof(1) and gprof(1)
, 169
programming errors
adding values with different
magnitudes
, 88
ill-conditioned computations
,
94
taking the difference of similar
values
, 87
testing floating-point values for
equality
, 84
truncation to an integer value
,
92
unintentional underflow
, 91
Puma performance analysis tool
,
169
Q
QNaN
, 45
quad-precision
effect on performance
, 191
format
, 36
maximum values, table
, 56
precision of
, 40
quiet NaN
, 45
R
range
of denormalized values, table
,
43
range errors
, 74
related documentation
, 16
remainder
invalid operation
, 58
operation
, 62, 66
remainder function
, 118
remquo function
, 120
representation, floating-point
limits of
, 40
summary
, 48
results of floating-point
computations, factors
affecting
, 69
architecture type
, 80
basic operations
, 71
binary/decimal conversions
, 76
checklist
, 217
compiler behavior and
compiler version
, 78
compiler options
, 79
exception conditions
, 74
floating-point coding practices
,
83
math library errors
, 74
math library functions
, 73
operating system release
, 81
status register values
, 82
system-related
, 76
rint function
, 119
round function
, 120
round to nearest
, 55
round to nearest integral value
operation
, 62
round toward negative infinity
,
55
round toward positive infinity
,
55
round toward zero
, 55
rounding
errors
, 71
inexact result condition and
,
53
rounding modes
examining and setting
, 130
IEEE
, 54
library functions and
, 55
optimization and
, 56