Nothing hard but our subjective evaluation running WRF and MM5 showed a
notable improvement in overall wall-clock time-to-completion.
gerry
Mark Hahn wrote:
DX38BT have a 82566DC Gigabit controller, that isn't capable of jumbo
frames
(frame size is limited to 1500 bytes).
But S3200SH have a 82566 controller too and a 82541PI that support jumbo
frames up to 16kb (probably your switch will support frames up to 10kb).
do you have any data showing that jumbo frames will make a significant
difference? yes, they will reduce interrupt-handling overhead (and perhaps
other overheads), but those tend to be fairly minor concerns on modern,
obscenely-fast cores...
_______________________________________________
Beowulf mailing list, [email protected]
To change your subscription (digest mode or unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf
--
Gerry Creager -- [EMAIL PROTECTED]
Texas Mesonet -- AATLT, Texas A&M University
Cell: 979.229.5301 Office: 979.458.4020 FAX: 979.862.3983
Office: 1700 Research Parkway Ste 160, TAMU, College Station, TX 77843
_______________________________________________
Beowulf mailing list, [email protected]
To change your subscription (digest mode or unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf
