PSTSWM T3E-900 Algorithm Comparison

Performance Studies using

PSTSWM

SGI/Cray Research T3E-900 Protocol Performance

(distributed 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:

Distributed LT (1) (mpi)
Algorithm Comparison
  T42L1     T21L2     T42L2     T85L2     T85L1     T85L4  
  P=32     P=16     P=8     P=32     P=16     P=8  
  optimal algorithm   halfsum  halfsum  ringpipe  halfsum  halfsum  ringpipe 
  (allreduce-min)/min     0.503    0.241    0.198    0.693    0.328    0.223 
  (generic-min)/min     0.509    0.354    0.065    0.160    0.079    0.070 

Distributed LT (1) (shmem)
Algorithm Comparison
  T42L1     T21L2     T42L2     T85L2     T85L1     T85L4  
  P=32     P=16     P=8     P=32     P=16     P=8  
  optimal algorithm   halfsum  halfsum  ringpipe  ringpipe  ringpipe  ringpipe 
  (generic-min)/min     0.600    0.443    0.230    0.438    0.279    0.109 

Distributed LT (1) (combined)
Communication Library Comparisons
  T42L1     T21L2     T42L2     T85L2     T85L1     T85L4  
  P=32     P=16     P=8     P=32     P=16     P=8  
  optimal library   shmem  shmem  shmem  shmem  shmem  shmem 
  (allreduce-min)/min     1.161    0.765    0.429    1.351    0.687    0.280 
  (mpi-min)/min     0.438    0.422    0.193    0.388    0.270    0.047 

Patrick H. Worley / ( worleyph@ornl.gov)
Last Modified Monday, 15-Jul-2002 10:24:40 EDT.