PSTSWM SP3-200 Communication Protocol Performance

Performance Studies using

PSTSWM

IBM SP3-200 Protocol Performance

(distributed LT experiment C / ring-pipeline sum algorithm)

Date/Person: June 8, 1999 / P. Worley
Platform: IBM SP3 at Oak Ridge National Laboratory (morgan.ccs.ornl.gov):
     62 2-way Winterhawk I SMP nodes (200 MHz POWER3 with 4MB L2 cache, equivalent to RS/6000 Model 260)
Environment: AIX 4.3.2;   PSSP 3.1
Code Version: 6.5
Compilation Options: mpxlf -O3 -qarch=auto -qtune=auto -qcache=auto
Link Options: -bmaxdata:0x70000000
Math Library: ESSL
Communication Library: MPI over the switch using user space for internode communication
MPI using IP for intranode communication
Parallel Algorithm: ringpipe
Number of Timesteps: 12
Results:

4x16 Processors (Column Major Ordering) / Problem T42L2
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  1.1973e-01   0.42   0.40   0.89 
Three Fastest
Protocols
1st2nd3rd
  i4   i5   g5 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  1   5   10 

2x32 Processors (Column Major Ordering) / Problem T42L2
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  1.3377e-01   0.69   0.63   1.84 
Three Fastest
Protocols
1st2nd3rd
  f4   f5   i5 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  2   2   8 

8x8 Processors (Column Major Ordering) / Problem T42L16
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  4.5376e-01   0.11   0.10   0.25 
Three Fastest
Protocols
1st2nd3rd
  f5   c4   i5 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  2   8   50 

4x16 Processors (Column Major Ordering) / Problem T85L4
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  4.4144e-01   0.23   0.21   0.61 
Three Fastest
Protocols
1st2nd3rd
  f5   f4   i4 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  1   5   33 

2x32 Processors (Column Major Ordering) / Problem T85L4
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  5.1551e-01   0.38   0.33   0.82 
Three Fastest
Protocols
1st2nd3rd
  i4   f4   f5 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  4   4   16 

8x8 Processors (Column Major Ordering) / Problem T85L32
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  3.1090e+00   0.09   0.08   0.23 
Three Fastest
Protocols
1st2nd3rd
  d4   f1   f3 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  2   13   50 

Patrick H. Worley / ( worleyph@ornl.gov)
Last Modified Monday, 15-Jul-2002 10:16:43 EDT.