Greg Lindahl wrote:
On Tue, May 30, 2006 at 09:12:28AM -0500, Richard Walsh wrote:

   2.  Three tiered memory system with simple local memory (store,
   like old Cray-2) that is user/software managed is preferable to
   cache in the above context.  Double buffering and prefetching to
   local store reduce memory delays dramatically.

I wouldn't make this comparison. The Cray 2 local store was much
easier to use than Cell's software-managed memory. And almost no one
actually used the Cray 2 local store. Which ought to be a lesson
learned.
OK then perhaps, >>more like<< the Cray-2's local memory than like a cache, and I used it (but I guess that definitely qualifies as "almost no one" ... ;-) ...). Are you wanting to imply that user controlled chip-local storage is not that great a design feature for HPC generally or more narrowly that the CELL's implementation could be better?
Another example of a modern machine with software-managed local store
is QCDOC. Again, this is the easy-to-use kind, very unlike Cell.

-- greg

_______________________________________________
Beowulf mailing list, [email protected]
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf



--

Richard B. Walsh

Project Manager
Network Computing Services, Inc.
Army High Performance Computing Research Center (AHPCRC)
[EMAIL PROTECTED]  |  612.337.3467

-----------------------------------------------------------------------
This message (including any attachments) may contain proprietary or
privileged information, the use and disclosure of which is legally
restricted.  If you have received this message in error please notify
the sender by reply message, do not otherwise distribute it, and delete
this message, with all of its contents, from your files.
-----------------------------------------------------------------------
_______________________________________________
Beowulf mailing list, [email protected]
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Reply via email to