Text item:
Now that the "Pentium Pro" is officially announced, are you able
to get (and release) STREAM numbers?
john
-- John D. McCalpin mccalpin@perelandra.cms.udel.edu Assistant Professor http://perelandra.cms.udel.edu/~mccalpin University of DelawareJohn
I have got some Pentium and Pentium Pro numbers. Pentium numbers were collected from Gateway 2000 P5-133/66 product, but Pentium Pro numbers were from test systems. I will get some numbers from OEM systems.
My Pentium numbers show some differences between Assignment and other operations. Anybody else has encountered this problem? The following two sets show the problem. I thought all four operations should have similar bandwidth numbers (?)
-- jason
(1) Stream 2.0
---------results from P5-133/66 256K PBSRAM, 64EDO DRAM, ------------ -- UNIXware, Intel Reference Compiler "icc -tp P5 -ip stream_d.c" ---
Timing calibration 1: utime overhead = 0.000000 usec. Timing calibration 2: array init time = 500000.000000 usec. Increase the size of the arrays if this is < 300000 and your clock precision is =< 1/100 second. --------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Assignment: 84.771 0.097 0.090 0.100 Scaling : 127.156 0.070 0.060 0.070 Summing : 127.157 0.097 0.090 0.100 SAXPYing : 114.441 0.103 0.100 0.110
(2) Stream 4.0 Beta
---------Result from P5-133/66 256K PBSRAM, 64EDO DRAM, -------------- ----- NT3.51, Visual C++ 2.2, code generation "Pentium" was selected--
------------------------------------------------------------- This system uses 8 bytes per DOUBLE PRECISION word. ------------------------------------------------------------- Array size = 2000000, Offset = 0 Total memory required = 45.8 MB. Each test is run 10 times, but only the *best* time for each is used. ------------------------------------------------------------- Your clock granularity/precision appears to be 9999 microseconds. Each test below will take on the order of 250000 microseconds. (= 25 clock ticks) Increase the size of the arrays if this shows that you are not getting at least 20 clock ticks per test. ------------------------------------------------------------- WARNING: The above is only a rough guideline. For best results, please be sure you know the precision of your system timer. ------------------------------------------------------------- Function Rate (MB/s) RMS time Min time Max time Assignment: 91.4286 0.3616 0.3500 0.3810 Scaling : 114.2857 0.2895 0.2800 0.3100 Summing : 125.9843 0.4042 0.3810 0.4510 SAXPYing : 114.0143 0.4357 0.4210 0.4600
Text item: External Message Header
The following mail header is for administrative use and may be ignored unless there are problems.
***IF THERE ARE PROBLEMS SAVE THESE HEADERS***.
Subject: Re: Re[6]: gcc compiler flags To: Jason J Ding <Jason_J_Ding@ccm.sc.intel.com> Message-Id: <9511031523.AA22038@perelandra.cms.udel.edu> From: mccalpin@perelandra.cms.udel.edu (John D. McCalpin) Date: Fri, 3 Nov 95 10:23:41 -0500 Received: by perelandra.cms.udel.edu (931110.SGI/930416.SGI) for @udel.edu:Jason_J_Ding@ccm.sc.intel.com id AA22038; Fri, 3 Nov 95 10:23 :41 -0500 Received: from perelandra.cms.udel.edu (mccalpin@perelandra.cms.udel.edu [128.17 5.74.1]) by copland.udel.edu (8.6.12/8.6.12) with SMTP id KAA18047 for <@udel.ed u:Jason_J_Ding@ccm.sc.intel.com>; Fri, 3 Nov 1995 10:24:04 -0500 Received: from copland.udel.edu by hermes.intel.com (5.65/10.0i); Fri, 3 Nov 95 07:30:37 -0800 Received: from hermes.intel.com by ccmgate1.sc.intel.com with smtp (Smail3.1.28.1 #3) id m0tBO0e-000S0XC; Fri, 3 Nov 95 07:26 WET
This archive was generated by hypermail 2b29 : Tue Apr 18 2000 - 05:23:05 CDT