HP Integrity rx8640 Stream results

Stream results for an HP Integrity rx8640 with the sx2000 chipset,
and 1.6GHz/6MB Itanium 2 processors:

4 cells, 16 processors, 128GB of memory (64x2GB DIMMs):
---------------------------------------------------
Function Rate (MB/s) Avg time Min time Max time
Copy: 29096.7522 0.0353 0.0352 0.0354
Scale: 28672.1380 0.0358 0.0357 0.0360
Add: 32219.5363 0.0479 0.0477 0.0483
Triad: 32472.2271 0.0480 0.0473 0.0523

2 cells, 8 processors, 64GB of memory (32x2GB DIMMs):
--------------------------------------------------
Function Rate (MB/s) Avg time Min time Max time
Copy: 14540.4956 0.0708 0.0704 0.0715
Scale: 14406.1301 0.0712 0.0711 0.0713
Add: 16187.6889 0.0950 0.0949 0.0952
Triad: 16271.2149 0.0946 0.0944 0.0952

The system was configured with half of the memory in each
cell assigned to local memory.

The runs used the v5.6 f90 version of the stream benchmark, with
the following changes:

97c97
< PARAMETER (n=2000000,offset=0,ndim=n+offset,ntimes=10)
---
> PARAMETER (n=64000064,offset=0,ndim=n+offset,ntimes=10)
124c124
< * COMMON a,b,c
---
> COMMON a,b,c
245c245
< 9020 FORMAT (1x,a,i4,a)
---
> 9020 FORMAT (1x,a,i6,a)

compiled as follows:

f90 -o stream_d.omp +Ofaster +DD64 +extend_source +autodbl4 +noppu
-Wl,+pd,1M +Oopenmp stream.f my second.o

and run with the mpsched -P RR command to distribute the processes in a
round robin
fashion across the locality domains. By default each thread allocated
memory
from the local memory in each cell. Here are the outputs for each
configuration:

4 cells, 16 processors, 128GB of memory (64x2GB DIMMs):

mpsched -T RR stream_d.omp
----------------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLE PRECISION word
----------------------------------------------
 ----------------------------------------------
 STREAM Version $Revision: 5.6 $
 ----------------------------------------------
 Array size = 64000064
 Offset = 0
 The total memory requirement is 1464 MB
 You are running each test 10 times
 --
 The *best* time for each test is used
 *EXCLUDING* the first and last iterations
 ----------------------------------------------
 Number of Threads = 16
 ----------------------------------------------
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 ----------------------------------------------------
 Your clock granularity/precision appears to be 1 microseconds
 ----------------------------------------------------
Function Rate (MB/s) Avg time Min time Max time
Copy: 29096.7522 0.0353 0.0352 0.0354
Scale: 28672.1380 0.0358 0.0357 0.0360
Add: 32219.5363 0.0479 0.0477 0.0483
Triad: 32472.2271 0.0480 0.0473 0.0523
 ----------------------------------------------------
 Solution Validates!
 ----------------------------------------------------

2 cells, 8 processors, 64GB of memory (32x2GB DIMMs):

mpsched -T RR stream_d.omp
----------------------------------------------
 Double precision appears to have 16 digits of accuracy
 Assuming 8 bytes per DOUBLE PRECISION word
----------------------------------------------
 ----------------------------------------------
 STREAM Version $Revision: 5.6 $
 ----------------------------------------------
 Array size = 64000064
 Offset = 0
 The total memory requirement is 1464 MB
 You are running each test 10 times
 --
 The *best* time for each test is used
 *EXCLUDING* the first and last iterations
 ----------------------------------------------
 Number of Threads = 8
 ----------------------------------------------
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 Printing one line per active thread....
 ----------------------------------------------------
 Your clock granularity/precision appears to be 1 microseconds
 ----------------------------------------------------
Function Rate (MB/s) Avg time Min time Max time
Copy: 14540.4956 0.0708 0.0704 0.0715
Scale: 14406.1301 0.0712 0.0711 0.0713
Add: 16187.6889 0.0950 0.0949 0.0952
Triad: 16271.2149 0.0946 0.0944 0.0952
 ----------------------------------------------------
 Solution Validates!
 ----------------------------------------------------




Received on Tue Mar 21 21:43:44 2006