Stream results for HP 9000 rp7420-16 with 1000MHz dual processor PA-8800
modules with 32MB external cache (shared by each processor pair):
System Copy Scale Add Triad
HP 9000 rp7420-16 1 cell, 8 cpus 4979 4971 4910 5098
HP 9000 rp7420-16 2 cells, 8 cpus 7877 7844 8672 8689
HP 9000 rp7420-16 2 cells, 16 cpus 8030 7964 8753 8809
The 2 cell configuration with 2 PA-8800 modules per cell
(half populated) is orderable and fully supported by HP.
The system was running HP-UX 11i TCOE (December 2003), with
all memory interleaved across the cells.
The f90 version of the stream benchmark was compiled auto-parallel, with
the following changes (mysecond.c is a C routine that calls gettimeofday):
63c63
< PARAMETER (n=2000000,offset=0,ndim=n+offset,ntimes=10)
--- > PARAMETER (n=288000320,offset=0,ndim=n+offset,ntimes=10) 72c72 < INTEGER bytes(4) --- > INTEGER*8 bytes(4) 90c90 < * COMMON a,b,c --- > COMMON a,b,c 200c200 < 9020 FORMAT (1x,a,i4,a) --- > 9020 FORMAT (1x,a,i8,a) rm -f *.o stream_d.mp stream_d.uni stream_c.mp stream_c.uni cc +DD64 +O3 -c mysecond.c f90 -o stream_d.mp +Ofaster -Wl,+pd,L +DD64 +Oautopar +Onoopenmp +autodbl4 +extend_source +noppu stream_d.f mysecond.o results for 1 cell, 8 processors (4 PA-8800 modules), 16GB of memory (16x1GB DIMMs): ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 288000320 Offset = 0 The total memory requirement is 6591 MB You are running each test 10 times -- The *best* time for each test is used *EXCLUDING* the first and last iterations ---------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds ---------------------------------------------------- Function Rate (MB/s) Avg time Min time Max time Copy: 4979.2749 0.9333 0.9254 0.9857 Scale: 4970.8471 0.9346 0.9270 0.9857 Add: 4909.6093 1.4092 1.4079 1.4104 Triad: 5098.1855 1.4015 1.3558 1.4086 ---------------------------------------------------- Solution Validates! ---------------------------------------------------- results for 2 cells, 8 processors (4 PA-8800 modules), 32GB of memory (32x1GB DIMMs): ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 288000320 Offset = 0 The total memory requirement is 6591 MB You are running each test 10 times -- The *best* time for each test is used *EXCLUDING* the first and last iterations ---------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds ---------------------------------------------------- Function Rate (MB/s) Avg time Min time Max time Copy: 7877.3744 0.5884 0.5850 0.5912 Scale: 7844.0006 0.5908 0.5875 0.5940 Add: 8671.5414 0.7998 0.7971 0.8047 Triad: 8689.1576 0.8168 0.7955 0.8752 ---------------------------------------------------- Solution Validates! ---------------------------------------------------- results for 2 cells, 16 processors (8 PA-8800 modules), 32GB of memory (16x512MB DIMMs): ---------------------------------------------- Double precision appears to have 16 digits of accuracy Assuming 8 bytes per DOUBLE PRECISION word ---------------------------------------------- Array size = 288000320 Offset = 0 The total memory requirement is 6591 MB You are running each test 10 times -- The *best* time for each test is used *EXCLUDING* the first and last iterations ---------------------------------------------------- Your clock granularity/precision appears to be 1 microseconds ---------------------------------------------------- Function Rate (MB/s) Avg time Min time Max time Copy: 8030.2633 0.5755 0.5738 0.5772 Scale: 7963.5097 0.5865 0.5786 0.6346 Add: 8753.1337 0.7909 0.7897 0.7923 Triad: 8808.9195 0.7905 0.7847 0.8122 ---------------------------------------------------- Solution Validates! ----------------------------------------------------Received on Tue Feb 10 16:21:19 2004
This archive was generated by hypermail 2.1.8 : Wed Feb 11 2004 - 16:09:54 CST