On Jan 22, 2004, at 15:45, Sean C. Garrick wrote:
> While the G5 is faster, the Opteron seems to have a better memory
> access/bandwidth.
Does your code access memory sequentially or randomly? The Opteron
should fare significantly better in the latter case thanks to its
on-chip memory controller - ~60ns for the Opteron vs 135ns for the G5.
If that doesn't account for much of the difference you might want to
play around with various compiler options regarding alignment and loop
unrolling - as a simple first pass you might try gcc's -fast option or
xlc's -qalign=natural -qunroll=(yes|auto).
Chris
[demime 0.98b removed an attachment of type application/pkcs7-signature which had a name of smime.p7s]
_______________________________________________
scitech mailing list | email@hidden
Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/scitech
Do not post admin requests to the list. They will be ignored.
References:
>G5 bandwidth (From: "Sean C. Garrick" <email@hidden>)
