PCMark05 Favors Intel CPUs ?

jmke · 30th July 2008, 16:35

I've heard rumors for years that performance in PCMark 2005 could change depending on what CPUID was handed to the benchmark, but this is the first opportunity I've ever had to test that theory. The term CPUID refers to a processor-specific character string that stores information on the chip's manufacturer, available features, make, and model. Different manufacturers use different CPUIDs, including GenuineIntel, AuthenticAMD, CentaurHauls, and the now-obsolete CyrixInstead. Intel and AMD both lock their CPUIDs to prevent them being changed by a third party, but VIA doesn't—and that gives us an opportunity to explore a question that normally can't be explored.

My my. Swap CentaurHauls for AuthenticAMD, and Nano's performance magically jumps about 10 percent. Swap for GenuineIntel, and memory performance goes up no less than 47.4 percent. This is not a test error or random occurance; I benchmarked each CPUID multiple times across multiple reboots on completely clean Windows XP installations. The gains themselves are not confined to a small group of tests within the memory subsystem evaluation, but stretch across the entire series of read/write tests.

http://arstechnica.com/reviews/hardw...o-review.ars/6

30th July 2008, 17:36

Hmmm....

Massman · 30th July 2008, 18:08

Interesting ... I really hope that other people pick up on this and re-run the tests as well :-).

jmke · 1st August 2008, 15:19

Slashdot posted this too here: http://hardware.slashdot.org/article.../08/01/1152237

with some interesting comments:

Quote:

I'll give 10:1 odds that Futuremark simply compiled their benchmark with Intel's C++ compiler.

I wrote a detailed explanation [slashdot.org] back in 2005 about how the Intel C++ compiler generates separate code paths for memory operations to make AMD processors appear significantly slower, and how you can trick the compiled code into believing your AMD processor is an Intel one to see incredibly increased performance.
http://yro.slashdot.org/article.pl?s...320202&tid=142

1st August 2008, 23:02

No matter what the cause, it means any benchmark that's not on an Intel system is flawed and any cross-platform comparisons are flawed.
Pretty pointless using the software unless all you want to do is to note incremental changes in a specific system after any upgrades?

30th July 2008, 16:35	#1
jmke Madshrimp Join Date: May 2002 Location: 7090/Belgium Posts: 79,021	PCMark05 Favors Intel CPUs ? I've heard rumors for years that performance in PCMark 2005 could change depending on what CPUID was handed to the benchmark, but this is the first opportunity I've ever had to test that theory. The term CPUID refers to a processor-specific character string that stores information on the chip's manufacturer, available features, make, and model. Different manufacturers use different CPUIDs, including GenuineIntel, AuthenticAMD, CentaurHauls, and the now-obsolete CyrixInstead. Intel and AMD both lock their CPUIDs to prevent them being changed by a third party, but VIA doesn't—and that gives us an opportunity to explore a question that normally can't be explored. My my. Swap CentaurHauls for AuthenticAMD, and Nano's performance magically jumps about 10 percent. Swap for GenuineIntel, and memory performance goes up no less than 47.4 percent. This is not a test error or random occurance; I benchmarked each CPUID multiple times across multiple reboots on completely clean Windows XP installations. The gains themselves are not confined to a small group of tests within the memory subsystem evaluation, but stretch across the entire series of read/write tests. http://arstechnica.com/reviews/hardw...o-review.ars/6 __________________ Last edited by jmke : 30th July 2008 at 16:37.

30th July 2008, 18:08	#3
Massman [M] Reviewer Join Date: Nov 2004 Location: Waregem Posts: 6,466	Interesting ... I really hope that other people pick up on this and re-run the tests as well :-). __________________

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Intel Talks "Canoe Lake" Platform and "Oak Trail" CPUs	jmke	WebNews	0	1st June 2010 15:08
88 percent of Intel CPUs in Q1 2011 to be Core 2010	jmke	WebNews	2	12th May 2010 22:41
Intel Plans Core i7 and Core i5 CPUs for Notebooks in 2010	jmke	WebNews	0	13th November 2009 13:00
Early sales of Core i5 and i7 CPUs unlikely to stop despite Intel warning	jmke	WebNews	1	21st August 2009 23:21
Intel Has Shipped 1 Million Nehalems CPUs	jmke	WebNews	2	17th April 2009 20:36
QPI Unlocked on recent Intel Core i7 920 and 940 CPUs	jmke	WebNews	4	7th March 2009 08:37
Intel Core i7 makes up for 1% of consumer CPUs	jmke	WebNews	0	10th February 2009 19:56
Noctua presents NH-U12P SE1366 for Intel Core i7 CPUs	jmke	WebNews	0	12th November 2008 20:46
Intel Nehalem to be released as Core i7 CPUs	jmke	WebNews	0	8th August 2008 15:42
Intel delivers 64-bit CPUs to server OEMs, delays shipments for the desktop	jmke	WebNews	0	12th August 2004 22:37

1st August 2008, 23:02	#5
FireTech Posts: n/a	No matter what the cause, it means any benchmark that's not on an Intel system is flawed and any cross-platform comparisons are flawed. Pretty pointless using the software unless all you want to do is to note incremental changes in a specific system after any upgrades?