Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: Joe NYC who wrote (76515)	4/4/2002 4:59:29 PM
From: wanna_bmw	Read Replies (1) \| Respond to of 275872

Joe, Re: "I think you mentioned that you attended. Did you hear anything from any him or some other AMD representative?" I remember the quote clearly. He said it after his presentation during a Q&A session in front of the audience. Someone asked about performance, and he said that Hammer could reach "roughly double the current SPEC maximum". Hans is correct that Webber did not mention frequency, but neither did he mention the processor mode (32-bit or 64-bit), the memory configuration (single channel or dual channel), the product proliferation (Clawhammer or Sledgehammer), or the basis for "maximum" (Power4 maximum, Athlon maximum, other??). I have my theory, based on context. For one thing, AMD would surely represent the score in spec derived from a 64-bit compile. This puts their product in the best possible light. Second, AMD was certainly talking about their "Sledgehammer" design, given that the whole presentation was centered around a multiprocessor implementation. At no point did Webber ever utter the words "Clawhammer" or "Sledgehammer". He only said "Hammer", which makes the reference ambiguous. However, under the context of the presentation, one can infer that he meant the multiprocessor, dual memory channel, large cache Sledgehammer design. Third, one should realize that IBM posted the SPEC scores of the Power4 very close to the time when the Microprocessor Forum started. Therefore, it's very probable that Webber did not know the SPEC score of the Power4 when answering the Q&A question. Putting together the pieces, I theorize that Webber estimates the SPEC score for a Sledgehammer processor (of unknown frequency) running in x86-64 mode to be as high as 1300 (double the score of the 2.0GHz Xeon score, the largest at the time), but not necessarily as high as 1300. Remember that saying "roughly double" leaves a lot of room to interpretation. To Fred Webber, 70% may be "roughly double". Realistically, I expect a desktop Clawhammer at 2.0GHz to reach 900-1000 in SPECint (running 32-bit). I reach this conclusion by taking the current maximum Athlon XP SPECint score, 697, add 20%, which is the difference from 1.67GHz to 2.0GHz, and then add another 20% for a realistic increase in micro-architectural scalability. This comes out to 1000, but then I scale back a little conservatively, because I know that SPEC doesn't scale perfectly with megahertz. wbmw