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.

Strategies & Market Trends : Gorilla and King Portfolio Candidates -- Ignore unavailable to you. Want to Upgrade?

To: Thomas Mercer-Hursh who wrote (33924)	10/27/2000 9:16:47 PM
From: Ali Chen	Respond to of 54805

Sorry, forgot to use fixed font, see next message.

To: Thomas Mercer-Hursh who wrote (33924)	10/27/2000 9:17:34 PM
From: Ali Chen	Read Replies (1) \| Respond to of 54805

Thomas, <One of my reactions is to think that Rambus is quite possibly a technology which only has only a small advantage at lower speeds, but has the ability to keep up with higher speeds, where other memory becomes an increasing bottleneck because it can't keep up ... but it is an explanation which would indicate why there was not much difference now, but that we were on the leading edge of a period where the difference would become far more material.> It is an interesting theory, and it is frequently raised time to time by Rambus supporters. Unfortunately, available data do not support it. To the contrary, if you take time to visit the Intel performance web sites, developer.intel.com developer.intel.com you may find the following data: BAPCO SYSmark2000/Win98: i820-RDRAM i815-SDRAM SDRAM advantage ------------------------------------------------ 800MHz 171 173 +2 933MHz 187 191 +4 ------------------------------------------------ Video*2000 MPEG-2/win98: i820-RDRAM i815-SDRAM SDRAM advantage ------------------------------------------------ 800MHz 26.37 27.65 +1.28 933MHz 28.92 30.33 +1.41 ------------------------------------------------ MultimediaMark99 - W98: i820-RDRAM i815-SDRAM SDRAM advantage ------------------------------------------------ 800MHz 2119 2137 +18 933MHz 2380 2412 +32 ------------------------------------------------ This means that the performance difference _increases_ with the growth of internal CPU frequency, and not decreases as the "high-speed benefit" theory would suggest. A quick regression analysis of benchmark scores published by Intel shows very good consistency of data as function of CPU core frequency. The trends have very clear asymptotics. They would tell you that the RDRAM-820 has approximately 25% disadvantage in performance as compared to SDRAM-133. Now I see why Intel hesitated to publish results for more scientific applications (SPEC numbers) on i815 SDRAM systems: the memory dependence in those benchmarks is bigger, therefore the Rambus would look like a real loser in this case. Sorry to break your theory:) - Ali