PSTSWM Paragon Communication Protocol Performance

Performance Studies using

PSTSWM

Intel Paragon Protocol Performance

(distributed LT experiment A1 / exchange sum algorithm)

Date/Person: May 13, 1998 / P. Worley
Platform: Intel Paragon XP/S 150 MP at Oak Ridge National Laboratory:
     1024 MP nodes (3 50-MHz iPSC/860 processors per node)
Environment: Paragon OSF/1 Release 1.0.4 Server 1.4 R1_4_5
f77/Paragon Paragon Version R5.0.3
Code Version: 6.3
Compilation Options: if77 -O4 -Mnodepchk -Knoieee -Msafealloc
Math Library: KAI
Communication Library: NX
Parallel Algorithm: exchsum
Partition: 1x8, 1x16, or 1x32
Results:

1x16 Processors / Problem T21L2
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  2.3676e-01   0.06   0.05   0.18 
Three Fastest
Protocols
1st2nd3rd
  d3   d2   d4 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  2   14   24 

1x32 Processors / Problem T42L1
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  4.5606e-01   0.08   0.06   0.22 
Three Fastest
Protocols
1st2nd3rd
  d3   d2   d5 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  3   7   24 

1x8 Processors / Problem T42L2
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  1.3030e+00   0.02   0.02   0.06 
Three Fastest
Protocols
1st2nd3rd
  d4   d6   b4 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  7   20   24 

1x16 Processors / Problem T85L1
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  2.4521e+00   0.03   0.02   0.07 
Three Fastest
Protocols
1st2nd3rd
  d5   d3   b5 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  11   18   24 

1x32 Processors / Problem T85L2
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  3.9959e+00   0.06   0.04   0.15 
Three Fastest
Protocols
1st2nd3rd
  b0   b1   b2 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  4   13   24 

1x8 Processors / Problem T85L4
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  1.4343e+01   0.03   0.03   0.05 
Three Fastest
Protocols
1st2nd3rd
  b5   b4   b6 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  3   24   24 

DISCUSSION

Partitions of the Paragon processor grid were used that match the processor subsets that the parallel algorithms would run on in a two dimensional data decomposition: 1x8, 1x16, and 1x32.
Patrick H. Worley / ( worleyph@ornl.gov)
Last Modified Monday, 15-Jul-2002 10:29:34 EDT.