AMD slams chip performance benchmarks

[Tweaker]

Chipmaker AMD has leveled criticisms against a widely-used benchmarking suite, calling it biased in the favor of its arch-rival Intel.

The move steps up AMD's efforts to ensure that its processors compare favorably with Intel's in the public eye, but some observers say both companies' rhetoric should be taken with a large pinch of salt.

In a presentation circulated to several hardware-oriented Web sites, AMD alleges that a recently released version of the Sysmark benchmarking tool, from industry consortium BAPCo, has been revised to downplay the strengths of AMD's Athlon XP processors, while emphasizing tests that portray Intel's Pentium 4 in a favorable light. AMD also confirmed that it joined BAPCo earlier this year, with the aim of influencing the way next year's Sysmark is formulated.

It's not that Intel is beating AMD. If a $600 chip from Intel can outperform a $200 chip from AMD, then great. Kudos to the Intel engineers, you know how to throw die size at the problem and pass your high costs back on to the consumer, who unfortunately doesn't know to look at anything other than the number before the "MHz" label.

The problem is that there is a *very* strong appearance that the criteria for selecting which tests to use and which to discard were based on whether or not the Intel CPU was strong in that area. That is a very poor way to design a benchmark!

90% of the Excel test is sorting??? What about calculation??? What about error checking or copy/paste??? Intel was strong in a single area, so the rest was discarded or made far, far less important.

The Excel test alone is enough to show that the new benchmark software was poorly designed.

You can read more about this here. (http://news.com.com/2100-1001-956079.html?tag=fd_top)

Source: C/NET News (http://news.com.com/2100-1001-956079.html?tag=fd_top)

[Kdr Kane]

After personally evaluating Sysmark 2002, last January, I found the testing suite insufficient for a variety of reasons.

I always recommend the suite of benchmarks from eTesting Labs (http://www.etestinglabs.com/). If it's not available for download, you can get a CD copy for $5.00.

[JANNA]

i totally disregard all benchmarking when selecting a product and i base my purchases on records of reliability and customer support. i know a lot of AMD fanboys are going to take that out of context. when i speak of reliability i am looking at it from more of an en-mass thing than a one person isolated example. after all there are yugos that run fine for hundreds of thousands of miles. AMD's major flaw in design is not in their CPU at all it has always been in the way that they implement the chipset architecture and their lack of care over quality control in regards to the motherboard makers that produce the boards and chipsets. if AMD took their chipset production and motherboard production and took as much care and attention to those areas as they do with their CPU i would be there with them. but u cant just produce a car body and then sub-contract out to the lowest bidder for the engines...

[taco_fox]

I believe I read somewhere that the only reason AMD makes chipsets is becuase they're the only ones who can bring out the full performace... even though other chipset makers seem to be passing them. AMD would rather get out of the chipset business.


On a side note, has anyone heard of the Tbreds cracking more since they have gone back to the rectangle design? I know the Palominos that had the square die were more resistant to cracking or chipping.