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 : Intel Corporation (INTC)
INTC 40.34-2.6%3:59 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: dale_laroy who wrote (162688)3/21/2002 4:02:58 PM
From: combjelly   of 186894
 
"This would also seem to confirm that Clawhammer will have 512KB L2 cache."

Yep, it would. So they are packing about 30 million additional transistors into 24mm^2, much of that would have to be regular structures like cache.

And this information gives us some interesting clues. According to sandpile.org, a Morgan has 25.2 million transistors. A Palomino has 37.5 million. That means the 192k of extra L2 cache is 12.3 million transistors, and that would imply that 256k would be 16.4 million. And that would mean that the extra HTT ports and wider memory bus on the SledgeHammer are only a few hundred thousand transistors, or they are already on the ClawHammer chip...

Or, turn it around. A ClawHammer with a 256k L2 cache would be around 50 million transistors, or 50% larger than a Palomino with the same amount of cache. If we strip away all of the L2 cache, we see that the Palomino has about 20 million transistors, which is close to the K7 Classic with 22 million, and the ClawHammer has 34 million. So the ClawHammer core has about 70% more transistors than a Palomino core. And that means it has got to have the same number of HTT ports or there are a lot more changes than have been discussed so far. Or both.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext