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.
Politics : Formerly About Advanced Micro Devices

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: kapkan4u who wrote (71639)9/10/1999 10:00:00 AM
From: Scot  Read Replies (2) of 1573827
 
Thread:

Apologies if this was already posted.

"First-Ever" Rambus Benchmarks....an excerpt

inqst.com

>>>>>>>>>begin quote<<<<<<<<<<
At the Intel Developer Forum (Sept 1st 1999 in Palm Springs, CA) Jay Bell, Senior Fellow of Dell Computer disclosed the first publicly released benchmark results for the Rambus/Camino platform.

Though the results have yet to be independently verified, performance figures quoted by Mr. Bell generally indicate that under normal conditions, Rambus degrades the performance of Microsoft Office 2000 applications by an average of 25% as compared to PC100. Using a benchmark called Office Bench, Dell's engineers contrasted a BX chipset platform using 100MHz SDRAM against a Camino platform with 800MHz Direct Rambus. Other than DRAM and chip set, both systems were configured identically with 500MHz processors.

On average, the three primary component applications of the Microsoft Office 2000 suite suffer a 25% performance degradation when running Direct RDRAM compared to PC100. This is in stark contrast to Intel's stated expectations of a positive performance impact due to the use of Direct RDRAM.



Office Bench with Concurrent Network Activity

As a fundamental point in his presentation, Mr. Bell pointed out that the hoped for performance advantage of Rambus in 3D games was not a key issue in Dell's market segment and end user requirements. Rather, network performance was much more significant.

Overall network throughput is not something that can be impacted by the CPU or DRAM performance of a client PC. Instead, the client performance issue will surface in Win2K enabled enterprises that program their servers to do weekly round-robin client backups over the network. When hit by the server for a backup (perhaps once a week), the client PC can experience a temporary performance degradation lasting several minutes while perhaps 50-100 MBytes of data is backed up to the server. Though there is a statistically low probability that this event will coincide with heavy usage of the client PC, Dell seems to believe that performance under these circumstances is worth benchmarking.

As a measure of the possible performance impact of Direct RDRAM under such a network load, Dell's engineers ran the Office Bench suite concurrently with a network backup task using LapLink software. With the LapLink network backup running in the background, the Office 2000 benchmark scores degraded significantly. But more importantly, the gap in performance between PC100 and RDRAM widened.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext