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 : Rambus (RMBS) - Eagle or Penguin -- Ignore unavailable to you. Want to Upgrade?

To: sylvester80 who wrote (54226)	9/20/2000 10:28:44 AM
From: Scumbria	Respond to of 93625

Sylvester, Does RDRAM have better latency than SDRAM when there are more than two transactions waiting to be serviced? Assuming all page hits for both cases. DRDRAM might achieve better latency for the subsequent transactions. I'm not sure why you are having a hard time understanding this, but the i815 vs. i820 benchmarks show that average latencies are longer on the i820 systems. Scumbria

To: sylvester80 who wrote (54226)	9/20/2000 10:47:47 AM
From: chic_hearne	Read Replies (1) \| Respond to of 93625

Re: As I've already posted before, current benchmarks do not do justice or measure the high bandwidth multithreaded multitasking net aware applications of today. Current benchmarks today are old (most of them were written during the single app DOS days) and measure old world apps. The world has moved on. I also see you didn't say anything about the benchmark links I posted. sylvester, Take off your rose colored glasses for a second. In the past Intel has BADLY wanted to get support for DRDRAM. Why didn't they do what you suggest? Why instead did they post benchmarks on their own site that made DRDRAM look inferior? Of course, none of the numbers matter because there is no serious ramp of DRDRAM coming from the Dramurai. chic

To: sylvester80 who wrote (54226)	9/20/2000 1:52:55 PM
From: Ali Chen	Read Replies (1) \| Respond to of 93625

<You have yet to post any specific data that disproves the dramreview technical data that RDRAM has better overall average latency than SDRAM.> Your lovely "dramreview", dramreview.com , says in concluding section themselves: "Memory system latency as seen by the processor is a complex function of the interaction of many different device and system parameters. Modeling these interactions accurately is difficult and imprecise, leaving actual measurement as the best method of understanding system performance and behavior..." This renders their perverted attempts to "prove" better Rambus latency as useless. Why? Because the best "actual measurement" would be an application-based benchmark. And it is a wide known fact that Rambus systems fails to regular SDRAM-133 system (i815) despite of vastly superior marketing "paramaters". <I will make the question very simple. Does RDRAM have better latency than SDRAM when there are more than two transactions waiting to be serviced? Yes or no?> And the answer is _when_? :) :) <As I've already posted before, current benchmarks do not do justice or measure the high bandwidth multithreaded multitasking net aware applications of today. Current benchmarks today are old (most of them were written during the single app DOS days) and measure old world apps. The world has moved on. > As _you_ posted? Really? The most popular BAPCO Sysmark2000 benchmark is a collection of interacting mostly multithreaded applications. These applications are well "of today's". They are simply most recent market applications. I guess you just made up another Rambus lie, which will go into Carl' list siliconinvestor.com under #m: all benchmarks are old.. ;) ;)