diff options
| author | Jeff Moyer <jmoyer@redhat.com> | 2009-10-02 18:57:36 -0400 |
|---|---|---|
| committer | Jens Axboe <jens.axboe@oracle.com> | 2009-10-28 09:29:25 +0100 |
| commit | cfb1e33eed48165763edc7a4a067cf5f74898d0b (patch) | |
| tree | d0e0bdd0664615b1f7be6cf770476e16dbcad116 /ipc | |
| parent | 1af60fbd759d31f565552fea315c2033947cfbe6 (diff) | |
| download | kernel_samsung_smdk4412-cfb1e33eed48165763edc7a4a067cf5f74898d0b.zip kernel_samsung_smdk4412-cfb1e33eed48165763edc7a4a067cf5f74898d0b.tar.gz kernel_samsung_smdk4412-cfb1e33eed48165763edc7a4a067cf5f74898d0b.tar.bz2 | |
aio: implement request batching
Hi,
Some workloads issue batches of small I/O, and the performance is poor
due to the call to blk_run_address_space for every single iocb. Nathan
Roberts pointed this out, and suggested that by deferring this call
until all I/Os in the iocb array are submitted to the block layer, we
can realize some impressive performance gains (up to 30% for sequential
4k reads in batches of 16).
Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Diffstat (limited to 'ipc')
0 files changed, 0 insertions, 0 deletions