AMD Processors
Decrease font size
Increase font size
Topic Title: PiFast 4.3 Fix 1 thread
Topic Summary:
Created On: 01/12/2004 01:52 PM
Status: Read Only
Linear : Threading : Single : Branch
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 01/12/2004 01:52 PM
User is offline View Users Profile Print this message

Author Icon
Senior Member

Posts: 3185
Joined: 12/07/2003

Alright, I got bored and made this thread to co-exist with the SuperPi thread.

I do a LOT of benchmarking here at the office when I get the chance, so if you'll pardon the lack of graphical evidence, I'm going to cut and paste my results from several machines here and at home (from a text file I've created, SO FANCY! ):

PiFast version 4.3 fix 1, times are for 1048576/4194504/10,000,000 places, all RAM disk, no swap file:

AMD XP1800+ @ 1715 MHz, 512MB DDR333, Windows 2000 Pro SP4 - 5.02 seconds / 27.67 seconds / 76.99 seconds.

Centrino 1.5 GHz 1MB L2, 512MB DDR333, Windows XP Pro SP1 - 4.14 seconds / 23.38 seconds / 66.08 seconds.

P4 Northwood mobile 2.0 GHz, 256MB DDR266, Windows 2000 Pro SP4 - 5.00 seconds / 27.32 seconds / 76.72 seconds.

P4 Willamette 1.6 GHz, 512MB SDRAM, Windows 2000 Pro SP4 - 7.27 seconds / 39.22 seconds / 107.81 seconds.

P4 Northwood "C" core 2.4GHz Hyper-Thread enabled, 512MB DDR333, Windows 2000 Pro SP4 - 5.20 seconds / 28.50 seconds / 78.33 seconds.

AMD XP 2500+ Barton @ 2250 MHz (12.5x180), 512MB DDR333, Windows 2000 Pro SP4 - 58.25 seconds for 10,000,000 places - only one I can remember right now. I'll re-run at home tonight and post a pic.

I've duplicated the weird P4 C core on 3 (count 'em: three) identical Dell boxes, and CPUID told me they were all what they were supposed to be. Very odd.

You can get PiFast here: http://numbers.computation.fre...stan...ram/pifast.html' ">http://numbers.computation.fre.../PiProgram/pifast.html

It's NOT multi-threaded, and doesn't recognize HT systems or SMP apparently. It's just a really good test of one CPU's FPU potential. [EDIT] - I might add that it DOES love memory bandwidth, so an Athlon FX-51 rig should mop up in this test. It's more power-hungry than SuperPi in that regard. Again, as mentioned in the SuperPi thread, it also uses a much better and quicker algorithm to compute Pi.

Post away!

112018 users are registered to the AMD Processors forum.
There are currently 0 users logged in.

FuseTalk Hosting Executive Plan v3.2 - © 1999-2015 FuseTalk Inc. All rights reserved.

Contact AMD Terms and Conditions ©2007 Advanced Micro Devices, Inc. Privacy Trademark information