PSTSWM Paragon Point-to-Point Communication Performance

Performance Studies using

PSTSWM


Intel Paragon SWAP Performance

(ordered swap of 8KB message using MPI)

Date/Person: January 30, 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
Communication Library: MPI
SWAP size: 1024 REAL*8 floating point values each direction
Message size: Largest - 1024 REAL*8 floating point values
Smallest - 1 REAL*8 floating point value
Processors: 0 and 1
Latency Definition:(T1024-T512)/512
Results:

ordered simple swap
Data Statistics
unidirectional bandwidth estimated latency model error
(peak MByte/sec) (usec/msg) (max. rel. error)
1 iter. 51.60 82.12 51.7%
10 iter. 60.58 82.88 61.3%

ordered swap using nonblocking send
Data Statistics
unidirectional bandwidth estimated latency model error
(peak MByte/sec) (usec/msg) (max. rel. error)
1 iter. 43.60 99.78 53.1%
10 iter. 49.34 100.19 60.4%
1 iter. w/overlap 38.57 116.74 55.0%
10 iter. w/overlap 45.41 107.06 59.4%

ordered swap using nonblocking receive
Data Statistics
unidirectional bandwidth estimated latency model error
(peak MByte/sec) (usec/msg) (max. rel. error)
1 iter. 43.44 99.95 53.0%
10 iter. 53.59 100.70 65.9%
1 iter. w/overlap 46.06 111.84 62.9%
10 iter. w/overlap 53.28 103.68 67.4%

ordered swap using nonblocking send and receive
Data Statistics
unidirectional bandwidth estimated latency model error
(peak MByte/sec) (usec/msg) (max. rel. error)
1 iter. 36.82 100.37 45.4%
10 iter. 47.62 100.38 58.3%
1 iter. w/overlap 37.18 143.40 65.1%
10 iter. w/overlap 46.02 129.33 72.7%

ordered swap using ready send
Data Statistics
unidirectional bandwidth estimated latency model error
(peak MByte/sec) (usec/msg) (max. rel. error)
1 iter. 28.23 150.86 52.0%
10 iter. 41.55 150.94 76.6%
1 iter. w/overlap 47.22 117.68 67.8%
10 iter. w/overlap 54.24 108.47 71.8%

ordered swap using nonblocking ready send
Data Statistics
unidirectional bandwidth estimated latency model error
(peak MByte/sec) (usec/msg) (max. rel. error)
1 iter. 34.20 149.33 62.4%
10 iter. 42.28 149.55 77.2%
1 iter. w/overlap 35.26 150.85 64.9%
10 iter. w/overlap 43.66 133.80 71.3%

synchronous
Data Statistics
unidirectional bandwidth estimated latency model error
(peak MByte/sec) (usec/msg) (max. rel. error)
1 iter. 41.34 116.51 58.8%
10 iter. 52.43 115.71 74.1%


Patrick H. Worley / ( worleyph@ornl.gov)
Last Modified Monday, 15-Jul-2002 10:29:19 EDT.
86833 accesses since 1/2/96.