Roland Dreier wrote:
Given this added complexity:
6 files changed, 226 insertions(+), 39 deletions(-)
and the unpleasantness of having if (BITS_PER_LONG == 64) various
places, can you quantify the improvement this gives?
Would it make more sense to do this for userspace first?
Its actually more important in user kernel since in user space MPI
coalescing solve most of the BW problems of small messages
In kernel modules this change improves the BW of small messages too.
And although it touching several files all of them are in the low level
driver thus not impact on stability of the core or ULPs.
Tziporet
_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
