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 212.97+3.3%12:25 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Joe NYC who wrote (73243)3/4/2002 5:59:10 PM
From: Ali ChenRead Replies (2) of 275872
 
Joe, "Based on your and wmbw's posts, I am changing my opinion on P4 L2 from most likely running at half (marketing) speed to most likely running at full speed."

Why changing too fast?

According to Intel technological papers on 0.18um
process, their 6-T SRAM cell cannot run above 900MHz.
It was two years ago. I guess some improvements may
make it to 1.1 GHz and slightly above. There is absolutely
no way that their current SRAM cell can run at 2.2GHz,
it is physically impossible for this process, and
equally unlikely for the new 0.13um.

Immediate inspection of die photographs on chiparchitect.com
shows that the L2 is arranged in two identical blocks.
It is fairly apparent that the L2 cache is
2-way-interleaved, each way is running at 1GHz clock,
and both ways are time-multiplexed making 2GHz of
aggregate bandwidth.

The whole discussion was again about physical clock
rate on different silicon areas in P4, not how fast
a tiny multiplexor can switch two data busses. It is
the same as in Rambus, where 8 banks of regular 100MHz
DRAM arrays are 8:1 multiplexed into a 800-MHz data
stream, while each bank simply cannot run at 800MHz
since there is no magic.

So, do not change your minds too fast.

Regards,
- Ali
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext