ralphc> I didn't try calling barrier() so I don't know the answer.
ralphc> When power is restored, I can try it. My guess is that
ralphc> it's a timing issue and not a code reordering issue.
Hmm, then we really better understand what's going on, because
otherwise you're just going to have trouble again if someone makes a
CPU with a faster mfence instruction...
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
