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 216.83+1.3%12:05 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Joe NYC who wrote (76509)4/4/2002 2:51:29 PM
From: combjellyRead Replies (2) of 275872
 
"it is in direction of benchmarks becoming a glorified stream benchmark."

But even in this scenario, ClawHammer will have it's advantages. A lot of attention has been paid to the whole chain, not only will latency to memory be better, so will latency to L2. Assuming I am reading the slides properly, it looks as if the latency to L2 will be 8 cycles as compared to 20 some odd for the current Athlons. If they went to that effort, they also likely made the L2 cache a true dual ported one and likely increased the bandwidth to the L2. With L1 cache evictions able to be run in parallel with cache reads, combined with a possibly improved pre-fetch and higher bandwidth to L2, then the data starvation that the current Athlon core has is going to be lessened. And that will have a big impact on SPEC, especially SPECfp. AMD was estimating a SPECint score of 1400 at 2GHz, so I suspect that Hammers will do just fine on STREAM-like benchmarks.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext