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 36.38-1.3%Dec 22 3:59 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Tenchusatsu who wrote (140400)7/29/2001 7:38:13 PM
From: Elmer  Read Replies (1) of 186894
 
Absolutely false. It doesn't matter whether you have a single multiprocessor bus or two P2P processor interfaces. They're all going to access one memory bus anyway, at least on Intel's and AMD's dual-CPU chipsets.

A point I have tried to make in the past. P2P processor interfaces only move the bottleneck someplace else, they don't eliminate it. Plus they add significantly to pin count and add at least another clock to snoop cycles. You know all this but perhaps others don't, on Intel's shared bus, each processor can snoop transactions as they take place while on a P2P interface the processors can't see the other's cycles directly. This adds complexity and delays. With large L2 caches dramatically cutting down on bus traffic and the extremely high bandwidth available to Intel's P4 bus, the only advantage to AMD's P2P interface was they didn't have to design it. After years of stumbling just trying to get the simplest 2-way P2P chipset working, the disadvantages are obvious.

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