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 (76518)	4/4/2002 6:46:17 PM
From: wanna_bmw	Read Replies (1) \| Respond to of 275872

Joe, Re: "L2 going from 256 to 512 added 13.8% to NW" This is perhaps the high end of the realizable gains that Northwood can achieve over Willamette, and I believe that the benefit was largely due to deficiencies in the original cache design. With 128 byte cache lines, the Pentium 4 cache assumes a lot of locality, which may not be present in a lot of applications. Thus, a larger cache freed up a large bottleneck, and I don't expect the same level of improvement in Hammer. Thus, on average, I still stand by my original estimate for a 20% increase in performance for Clawhammer relative to the Athlon XP at the same clock frequency. Results are going to vary, especially with apps that can take advantage of Hammer's SSE-2 instructions, etc, but I think that overall, 20% is a realistic number. Re: "there is no reason for AMD not to present the CPU in 64bit long mode, with extra general purpose registers. This can amount to performance increase of 5% to 7.5%" In that case, expect a SPECint of 945-1075, as long as AMD can get their compilers to turn out high performance code by the Hammer launch. At this point, I am skeptical that the compilers will just fall into place, so I expect initial measurements to made running with 32-bit code. Re: "- initial clock speed may reach 2.2 GHz on Clawhammer part" This is not unreasonable, but you should make it clear that there is no evidence backing up estimates for the Hammer launch frequency. Earlier this week, I theorized that AMD is delaying Thoroughbred because their .13u process is not binning out high frequencies at this point in time, and some recent evidence seems to back that up. Also, rumor has it that Hammer's first stepping was at 1.4GHz. There is certainly a lot of time and room for improvement, but I would not count on anything just yet. wbmw