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)
AMD 237.62-7.2%Nov 6 3:59 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: dougSF30 who wrote (40887)5/22/2001 3:19:15 PM
From: dale_laroyRead Replies (1) of 275872
 
>"But *some* P4 benches, like that older Anand one, show a smooth curve.

So, as someone on here (you?) suggested earlier, perhaps a stepping screwed something up?"<

It might have to do with one or more of several factors, almost all involving granularity.

The relatively smooth curves could result from the granularity of the plotting points. A plot of only arrays that are powers of two would look significantly different than a plot of arrays in 4KB sized steppings.

The anomalies could result from thrashing of the 8KB L1 cache at particular data set sizes, and could be influenced by fragmentation of the memory map by the PMMU, a situation that might be increased in a DP system.

The anomalies could be cause by voltage fluctuations resulting from improper traces on the system board or an overloaded power supply. I remember that the first prototype board we produced for the 25 MHz AMD29K rang so severely that what started out as a +5v/-5v clock signal was at +12.5v/-9v by the time it got to the processor. If such an effect were to occur, even to a lesser extent, in the P4, the result could be a clock signal at a couple of tenth of a higher voltage than nominal.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext