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 207.67+2.2%3:59 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: wanna_bmw who wrote (78337)4/25/2002 8:47:16 AM
From: fyodor_Read Replies (1) of 275872
 
Wanna: It sounds like you're saying that the 25% IPC number is only for the most conservative, non-memory intensive benchmarks, and that it should be much higher in other cases.

No, that's not what I was trying to convey.

My *guess* would be that the 25% is based purely on SPECint.

*If* this is indeed the case, trying to gauge any increase in performance in specific applications, e.g. Lightwave, using this figure would be an exercise in futility.

What I was trying to point out was simply that most of the benchmarks you listed have fairly well-known characteristics in terms how performance varies with memory bandwidth, cache size and, to some extent, memory latency. I.e. it would like be a better estimate of Hammer performance simply to apply those changes, which we know of - rather than to use AMD's generic (or perhaps very specific, if SPECint-based) 25% IPC increase.

I haven't run the numbers, so I don't know if this puts the Hammer in a worse light or in a better light.

-fyo
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext