Running lmbench on quad xeon with Andi's patch shows a reduction
in select latency in all cases compared to -pre10.
File select - times in microseconds - smaller is better
Average of 25 runs. -ak-axboe is select/poll patch +
blockhighmem.
select select select select select select select select
kernel 10 fd 100 fd 250 fd 500 fd 10 TCP 100TCP 250TCP 500TCP
--------------------- ------ ------ ------ ------ ------ ------ ------ ------
2.4.19-pre10 4.73 30.66 74.62 143.77 5.67 40.879 97.914 193.10
2.4.19-pre10-ak-axboe 3.85 27.83 71.55 142.78 4.86 37.517 94.205 191.54
2.4.19-pre10-aa4 4.54 29.48 73.96 147.52 5.60 39.485 98.895 194.58
2.4.19-pre10-ac2 4.63 29.60 73.33 141.86 5.64 39.328 98.898 192.82
-- Randy Hron http://home.earthlink.net/~rwhron/bigbox.html- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/