From: Kirby L. Collins (kcollins@rsn.hp.com)
Date: Mon Nov 08 2004 - 11:08:14 CST
Below are single cpu stream results for an HP Integrity rx1620-2, configured as follows:
HP Integrity rx1620-2
2x1.6GHz/3M Itanium 2 processors
16GB memory (8x2GB DIMMs)
HP-UX 11.23.09
Note that on an rx1620 with the 1.6GHz/3M processors the front-side-bus runs at 533MHz (267MHz DDR) instead of 400MHz.
----------------------------------------------
Double precision appears to have 16 digits of accuracy
Assuming 8 bytes per DOUBLE PRECISION word
----------------------------------------------
Array size = 32000960
Offset = 0
The total memory requirement is 732 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: 4382.1520 0.1211 0.1168 0.1497
Scale: 4393.3939 0.1167 0.1165 0.1172
Add: 5027.4792 0.1529 0.1528 0.1530
Triad: 5031.4290 0.1528 0.1526 0.1530
----------------------------------------------------
Solution Validates!
----------------------------------------------------
I used Revision 5.0 of the stream code, modified to set the array size and to place the a,b,c arrays in common:
63c63
< PARAMETER (n=2000000,offset=0,ndim=n+offset,ntimes=10)
--- > PARAMETER (n=32000960,offset=0,ndim=n+offset,ntimes=10) 88c88 < * COMMON a,b,c --- > COMMON a,b,cand compiled with HP Fortran 90 :
f90 -o stream_d.uni +O3 +DSitanium2 +DD64 +Onoopenmp +extend_source +autodbl4 +noppu -Wl,+pd,16M stream_d.uni.f second_wall.o
This archive was generated by hypermail 2.1.4 : Tue Nov 16 2004 - 07:37:25 CST