Parallel Programming Guide for HP-UX Systems

Index
214
LOOP_PRIVATE directive and pragma, 109,
111
arrays, 116
loop-carried dependences, 181, 187
loops
DO (Fortran), 111
exploit parallel code, 145
for (C), 111
induction variables in parallel, 113
parallelizing inner, 193
parallelizing outer, 192
privatizing, 108
safely parallelizing, 170
that manipulate variables, 108
triangular, 191
LSIZE, 180
M
malloc, 176
man pages, xvii
manual
parallelization, 109
synchronization, 109, 156
memory
hypernode local, 123
space, occupying same, 170
memory class, 109, 128
assignments, 127
declarations (C/C++), 127
declarations (Fortran), 127
misused, 168
node_private, 123, 126, 131
thread_private, 123, 126
memory_class_malloc, 176
message signature analysis, 58
message-passing, 7
misused
directives and pragmas, 187
memory classes, 168
monospace, xvi
MP_GANG, 32
MPI, 7
MPI object-space corruption, 58
MPI_MT_FLAGS, 34
multiple
data dependences, 134
multiple buffer writes detection, 58
N
nested
parallelism, 135
NO_LOOP_DEPENDENCE directive and
pragma, 189
node_private
example, 131
static assignment of, 129, 131
virtual memory class, 123, 126
nondeterminism of parallel execution, 187,
190
nonordered
dependences, 145
notational conventions, xvi
O
objects, stack-based, 128
offset indexes, 180
optimization
levels, 169, 203
ordered
data dependences, 134
parallelism, 145
sections, 147
ordered sections
hidden, 187
using, 150
ORDERED_SECTION directive and
pragma, 147
overflowing trip counts, 201
P
parallel
construct, 145
execution, 190
structure, 135
PARALLEL_PRIVATE directive and
pragma, 109, 120
parallelism
nested, 135
ordered, 145
thread, 135
parallelization
manual, 109
overhead, 193
parallelizing