SI
SI
discoversearch

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: Cirruslvr who wrote (10025)9/23/2000 3:16:21 PM
From: ptannerRead Replies (1) | Respond to of 275872
 
Cirruslvr, Re: "The 840 has double the memory bandwidth of the 820"

Thanks for digging up SPEC comparisons for 820/840. Would there also be a difference for the capacity of the FSB since P4 will have 4x100 vs ??? presumably much lower for 820/840.

-PT



To: Cirruslvr who wrote (10025)9/23/2000 6:30:05 PM
From: andreas_wonischRespond to of 275872
 
Cirrus, thanks for the comparison table. Looks like SPECfp2000 is a very bandwidth intense benchmark. If a P4 with 3.2 GB/sec memory bandwidth and a SSE2 optimized compiler can only gain ca. 10% on an IPC level compared to a P3 in this benchmark suite (if I recall the leaked SPECfp number correctly) this doesn't bode well for legacy FP performance. In fact, it must be terrible. (Scumbria won?) It will be interesting to see how the press will treat this discrepancy in optimized benchmark and real-world performance.

Andreas



To: Cirruslvr who wrote (10025)9/24/2000 2:01:33 AM
From: PetzRespond to of 275872
 
Cirruslvr, thanks for digging out the i840 vs i820 performance on each specFP benchmark, that was a lot of work! SWIM performance improvement on Willy/850 vs CuMine/i840 may be more than the 21% improvement of CuMine/i840 vs. CuMine i820. This would be because the FSB of PIII only has 2/3 the theoretical bandwidth of the i840 memory bandwidth but FSB of Willy is 3 times faster.

Petz