Whamcloud - gitweb
LU-13799 osc: Don't get time for each page
Getting the time when each batch of pages starts is
sufficiently accurate, and ktime_get() is several % of the
CPU time when doing AIO + DIO.
This relies on previous patches in this series.
Measuring this in milliseconds/gigabyte lets us measure the
improvement in absolute terms, rather than just relative
terms.
This patch reduces i/o time in ms/GiB by:
Write: 17 ms/GiB
Read: 6 ms/GiB
Totals:
Write: 237 ms/GiB
Read: 223 ms/GiB
IOR:
mpirun -np 1 $IOR -w -r -t 64M -b 64G -o ./iorfile --posix.odirect
Without the patch:
write 4030 MiB/s
read 4468 MiB/s
With patch:
write 4326 MiB/s
read 4587 MiB/s
Signed-off-by: Patrick Farrell <farr0186@gmail.com>
Change-Id: I02897bf810683bc77a7d09156cdb83ba1d25ebf1
Reviewed-on: https://review.whamcloud.com/39437
Reviewed-by: Wang Shilong <wshilong@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>