Replies to post #2477 on Advanced Micro Devices Inc (AMD)

[UpNDown]

wbmw, on model numbering

What is new here is that the direct comparisons to Intel frequencies is explicitly made. At least I believe this to be something AMD has not been explicit about before. The earliest model number correspondences were supposedly made against earlier AMD models.

[Not a Short]

So what is new here? Haven't we known this since the beginning? We already know from reviews that AMD's model numbers cannot stand up to Intel's frequencies. They are too aggressive, and only getting worse.

I'm not in the mood to spend time looking up old posts on SI to refresh your memory on what everyone said in the past. I'll just give this quick link to an article from 2001 that explains the original stance on the ratings:

http://www.theinquirer.net/?article=1207

"The so-called "PR Ratings" compare the performance of AMD Athlons with Palomino cores against their own processors and are based on 35 different benchmarks, according to a representative at the launch of the chips in Milano today."

[UpNDown]

wbmw, on processor benchmark(et)ing

We already know from reviews that AMD's model numbers cannot stand up to Intel's frequencies. They are too aggressive, and only getting worse.

The problem is, of course, exactly which benchmarks are run, and how. Exactly which of Intel's frequencies, etc. For instance, I think you'll find this story from The Register interesting:
http://www.theregister.co.uk/content/3/30300.html

Seems Intel has MobileMark 2002 results up for the Centrino, showing how it beats the 2.4 GHz Penium 4-M all hollow, 189 to 164.
http://www.intel.com/products/benchmarks/notebook/index.htm

But then, Intel releases a 2.5 GHz Pentium 4-M and, wonder of wonders, shows that on MobileMark 2002 the 2.2 GHz 4-M scores 181 and the 2.5 scores 193.
http://www.intel.com/products/benchmarks/notebook/business.htm

Hard for AMD to track such a moving target.