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 214.18-0.5%Dec 31 3:59 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: wanna_bmw who wrote (80808)5/28/2002 11:26:42 AM
From: Ali ChenRead Replies (1) of 275872
 
"In the first situation, they are saying that an inbound memory write needs to flush those cache lines from the CPU caches, which is correct, except that they forget to mention that modified data needs to be written back"

Wrong.
Inbounded memory Write does not need to flash cache lines
since it makes no sense in first place - the flushed
data will be over-written anyway. The only thing a CPU
does it marks the corresponding cache line as "invalid",
so when the CPU needs the same data, the BUI will read
them from memory.

The cacheline flush has to be done when "other agent" READs
from memory. In case when the CPU contained a newer,
"modified" image of the main memory location, the
memory must be updated (chache line flushed) before the
READ can proceed to the requesting agent.

The cache coherency protocol is one of the most important
part of SMP architecture, and it is very unprofessional
to mess up with bus snooping basics.

Instead of expressing your own (and wrong) opinions, I strongly suggest to familiarize yourself at least with popular
interpretation of SMP architecture like T.Shanley's
"PentiumPro and Pentium II System Architecture",
ISBN 0-201-30973-4, Addison-Wesley.

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