PSTSWM SP3-375 Communication Protocol Performance

Performance Studies using

PSTSWM

IBM SP3-375 Winterhawk II Protocol Performance

(distributed FFT experiment A)

Date/Person: December 12, 1999 / P. Worley
Platform: IBM SP3 at Oak Ridge National Laboratory (bobcat.ccs.ornl.gov):
     62 4-way Winterhawk II SMP nodes (375 MHz POWER3 with 8MB L2 cache)
Environment: AIX 4.3.3;   PSSP 3.1.1
Code Version: 6.6.4
Compilation Options: mpxlf -O3 -qarch=auto -qtune=auto -qcache=auto
Link Options: -bmaxdata:0x70000000
Environment Variables: export MP_SHARED_MEMORY="yes"
Math Library: ESSL
Communication Library: MPI over the switch using user space for internode communication
MPI using shared memory for intranode communication
Parallel Algorithm: distributed FFT
Number of Timesteps: 12
Results:

16x1 Processors / Problem T21L2
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  4.5792e-02   0.41   0.47   0.92 
Three Fastest
Protocols
1st2nd3rd
  c2   c0   c1 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  3   4   8 

32x1 Processors / Problem T42L1
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  7.3210e-02   0.28   0.27   0.52 
Three Fastest
Protocols
1st2nd3rd
  a5   a4   b4 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  2   2   12 

8x1 Processors / Problem T42L2
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  2.2027e-01   0.13   0.14   0.25 
Three Fastest
Protocols
1st2nd3rd
  c3   c5   c4 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  1   4   26 

16x1 Processors / Problem T85L1
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  3.1272e-01   0.12   0.12   0.23 
Three Fastest
Protocols
1st2nd3rd
  c5   d5   c3 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  1   5   26 

32x1 Processors / Problem T85L2
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  3.8964e-01   0.11   0.14   0.22 
Three Fastest
Protocols
1st2nd3rd
  c5   d5   c4 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  1   8   26 

8x1 Processors / Problem T85L4
Runtime Statistics
min(mean-min)/min(median-min)/min(max-min)/min
  1.7090e+00   0.14   0.11   0.43 
Three Fastest
Protocols
1st2nd3rd
  c3   c5   c4 
       Number of Proctocols With
Runtimes Within X% of Min
1%5%25%
  1   2   24 

Patrick H. Worley / ( worleyph@ornl.gov)
Last Modified Monday, 15-Jul-2002 10:19:02 EDT.