. PSTSWM T3E-900 Algorithm Comparison

Performance Studies using

PSTSWM

SGI/Cray Research T3E-900 Algorithm Comparison

(transpose LT experiment I-A )

Date/Person: May 11, 1998 / P. Worley
Platform: T3E-900 at National Energy Research Scientific Computing Center (mcurie.nersc.gov):
   532 450-MHz DEC Alpha EV5 RISC processors
Environment: UNICOS/mk 2.0.2.28
mpt 1.2.0.2
Cray CF90 Version 3.0.2.1
Code Version: 6.3
Compilation Options: -dp -Oscalar3 -Wl,-Dstreams=on
Math Library: libsci
Communication Library: MPI
SHMEM
Number of Timesteps: 12
Results:

Transpose LT (1) (mpi)
Algorithm Comparison
  T10L16    T21L8     T21L32    T21L16    T42L16 
  P=16     P=8     P=32     P=16     P=8  
  optimal algorithm   alltoallv  alltoallv  alltoallv  alltoallv  alltoallv 
  (alltoallv-min)/min     0.000    0.000    0.000    0.000    0.000 
  (generic-min)/min     0.236    0.118    0.171    0.111    0.022 

Transpose LT (1) (shmem)
Algorithm Comparison
  T10L16    T21L8     T21L32    T21L16    T42L16 
  P=16     P=8     P=32     P=16     P=8  
  optimal algorithm   swtrans  srtrans  swtrans  srtrans  srtrans 
  (generic-min)/min     0.156    0.093    0.104    0.083    0.032 

Transpose LT (1) (combined)
Communication Library Comparisons
  T10L16    T21L8     T21L32    T21L16    T42L16 
  P=16     P=8     P=32     P=16     P=8  
  optimal library   shmem  shmem  shmem  shmem  shmem 
  (alltoallv-min)/min     0.545    0.123    0.346    0.244    0.052 
  (mpi-min)/min     0.545    0.123    0.346    0.244    0.052 

Patrick H. Worley / ( worleyph@ornl.gov)
Last Modified Monday, 15-Jul-2002 10:28:50 EDT.