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 38.80+1.7%10:38 AM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Tenchusatsu who wrote (126875)2/7/2001 10:25:41 PM
From: Dan3  Read Replies (2) of 186894
 
Re: Wrong.

PCI on entry level servers is 4 bytes wide and runs at 33MHZ. PC133 is 8 bytes wide and runs at 133MHZ.

PCI on high end SHV servers is 8 bytes wide and runs at 66MHZ. PC133 on high end servers is interleaved so it's 16 or 32 bytes wide and still runs at 133MHZ. Neither SDRAM nor PCI are 100% efficient in terms of saturating their respective busses.

Low end i/o is 33 x 4 = 132 vs SDRAM at 133 x 8 = 1064

High end i/o is 66 x 8 = 528 vs SDRAM at 133 x 16 or 32 = 2128 or 4256.

Those calculations are MB/sec bandwidth.

And if P4 isn't positively drowning its memory bus with prefetch traffic, then why is dual channel PC800 necessary to equal the performance of a PIII or Athlon with single channel PC133? Why can't even dual channel PC600 supply enough bandwidth?

I'd be interested in an informed criticism but if all you can come up with is "Wrong" then never mind.

Dan

PS - I agree with you that SMP servers need more memory bandwidth, but it's for multitasking logic operations not for streaming i/o, and P4 in its present configuration looks like a disaster for SMP. P4 requires the equivalent of a high end server's multichannel memory subsystem to run Office! How can it do anything but fall on its face running high bandwidth server apps? Which may well explain why there is no P4 SMP.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext