We went through all 16 Super TSO patches to see which one causes the
performance degradation (relative to the original TSO implementation)
that was observed earlier, and it appears to be the last patch #16;
results below.
We can either provide a remote to the setup, or test incremental patches
if #16 can be broken in smaller pieces.
------------------------------------------------------------------------
--------------------------------------------------
Configuration:
4 way Opteron system .247 as TX system with kernel as in the test
result.
4 way Opteron system .226 as Rx system with kernel 2.6.12.2, Xframe I
NIC driver REL_2-0-1-6_LX installed on both systems, Mtu is set to 9000
on both systems.
Systems are connected back to back.
Run 8 nttcp connections from Tx system to Rx system for 60 seconds.
Test result:
OS on Tx system TSO Throughput CPU idle
--------------- ------ ----------- --------
2.6.5-7.135-smp ON 6.63Gb/s 59%
OFF 6.01Gb/s 47%
2.6.12.1 ON 6.63Gb/s 59%
OFF 6.12Gb/s 50%
2.6.13-rc2-mm2 ON 5.95Gb/s 57%
OFF 6.16Gb/s 50%
2.6.13-rc3-mm1 ON 5.96Gb/s 58%
OFF 6.16Gb/s 49%
2.6.13-rc3-mm1 ON 6.62Gb/s 56%
(superTSO reversed)OFF 6.16Gb/s 48%
2.6.13-rc1 ON 6.63Gb/s 57%
(superTSO 1-15) OFF 6.17Gb/s 50%
2.6.13-rc1 ON 5.93Gb/s 58%
(superTSO 1-16) OFF 6.16Gb/s 52%
Note: 2.6.5-7.135-smp is SuSE shipping kernel.
2.6.13-rc2-mm2 has tso patch 16.
2.6.13-rc3-mm1 has latest tso patch.
2.6.13-rc1 is from linux-2.6-before-super-tso.tar.gz.
Kim Shaw
408-366-4616
Test Engineer
Neterion
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
