Index

Symbols

2-by-2 switches

A

Ada	array and ring topology arrival phase

B

bandwidth barrier bidirectional data transfer bisection bandwidth bisection width

blocking communication, [1] blocking receive blocking send broadcast butterfly barrier

C

circuit switching cluster collective communication, [1], [2], [3] communication granularity

communication latency compiler directive critical section, [1], [2] cross over crossbar switch

D

deadlock, [1] departure phase diameter of network distributed memory multicomputer

divide and conquer domain decomposition dynamic network topology dynamic work load balancing

E

efficiency

ethernet

F

fan out broadcast

flops functional decomposition, [1]

G

gather

H

heap	hypercube network

I

immediate receive

immediate send

J

job queue

job scheduling

L

latency

livelock

M

manager/worker model Markov chain master construct matrix and torus topology matrix-matrix multiplication MCA memory allocation message latency MIMD MISD Monte Carlo MPI_Bcast MPI_Comm_rank MPI_Comm_size MPI_COMM_WORLD MPI_Finalize MPI_GATHER

MPI_Init MPI_Iprobe MPI_IRECV MPI_ISEND MPI_RECV MPI_REDUCE MPI_SCATTER MPI_SEND MPI_Sendrecv MPI_SUM MPI_TEST MPI_WAIT MPI_Wtime Multiple Instruction Multiple Data stream Multiple Instruction Single Data stream multiprocessing module multistage network mutex

N

network latency

nonblocking communication, [1] numerical integration

O

Omega interconnection network omp_get_thread_num()

omp_set_num_threads() overhead

P

packet switching parallel construct parallel region pass through performance cliff pipeline pipeline cycle pipeline latency

pleasingly parallel, [1] point-to-point communication, [1] POSIX prefix sum algorithm private clause process, [1] pseudorandom numbers Python

Q

quad double, [1]

quadrature quality up

R

reduction algorithm release phase

Romberg integration routing algorithm

S

scalability scaled speedup scatter shared clause shared memory multicomputer SIMD Simpson rule SIMT single construct Single Instruction Multiple Data stream single instruction multiple thread Single Instruction Single Data stream

Single Program Multiple Data stream Single Program, Multiple Data SISD space-time diagram speedup SPMD, [1] square partition stack static work load assignment strip partition supercomputer superlinear speedup switches

T

tasking thread, [1] thread divergence thread safe trapezoidal rule

trapping phase tree barrier tree network type 1 pipeline, [1] type 2 pipeline type 3 pipeline

W

work crew