* Re: [PATCHSET v2] block: IO polling improvements
@ 2016-11-07 15:28 Stephen Bates
2016-11-07 15:29 ` Jens Axboe
0 siblings, 1 reply; 3+ messages in thread
From: Stephen Bates @ 2016-11-07 15:28 UTC (permalink / raw)
To: linux-kernel, linux-block, axboe
> Fixed a few bugs in this, and addressed some review comments. Patches
> are against my 4.10 block branch, for-4.10/block.
Jens
Thanks for proposing this. Looks very cool. I will try and get you a
review and some testing this week...
Cheers
Stephen
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCHSET v2] block: IO polling improvements
2016-11-07 15:28 [PATCHSET v2] block: IO polling improvements Stephen Bates
@ 2016-11-07 15:29 ` Jens Axboe
0 siblings, 0 replies; 3+ messages in thread
From: Jens Axboe @ 2016-11-07 15:29 UTC (permalink / raw)
To: 1478202307-1947-1-git-send-email-axboe, linux-kernel, linux-block
On 11/07/2016 08:28 AM, Stephen Bates wrote:
>> Fixed a few bugs in this, and addressed some review comments. Patches
>> are against my 4.10 block branch, for-4.10/block.
>
> Jens
>
> Thanks for proposing this. Looks very cool. I will try and get you a
> review and some testing this week...
Sounds good. Be sure to use the latest git branch (for-4.10/dio),
there's been some fixes since the posting.
--
Jens Axboe
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCHSET v2] block: IO polling improvements
@ 2016-11-03 19:45 Jens Axboe
0 siblings, 0 replies; 3+ messages in thread
From: Jens Axboe @ 2016-11-03 19:45 UTC (permalink / raw)
To: axboe, linux-kernel, linux-block; +Cc: hch
Fixed a few bugs in this, and addressed some review comments. Patches
are against my 4.10 block branch, for-4.10/block.
Some performance results, running random reads against a fast device,
using either IRQ driven completions, classic polled, or the new hybrid
polled:
Method IOPS CPU% IOPS/CPU%
=========================================================
IRQ 476K 48.45% 9843
Poll 897K 99.9% 8982
Hybrid 836K 68.1% 12273
Classic polling is still a bit faster, but is uses a lot more CPU. The
rightmost column is an efficiency metric, and with the new hybrid
scheme, we're both faster and more efficient than IRQ driven. We should
be able to close the speed gap as well, so you can have your cake and
eat it too.
I've adjusted the default, so that classic polling is the default poll
method. To enable hybrid polling for a device, you would do:
# echo 0 > /sys/block/<dev>/queue/io_poll_delay
Also see the last patch for the other values this can be set to.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2016-11-07 15:52 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-11-07 15:28 [PATCHSET v2] block: IO polling improvements Stephen Bates
2016-11-07 15:29 ` Jens Axboe
-- strict thread matches above, loose matches on Subject: below --
2016-11-03 19:45 Jens Axboe
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).