From: Dave Chinner <david@fromorbit.com>
To: John Hubbard <jhubbard@nvidia.com>
Cc: Jan Kara <jack@suse.cz>, Christoph Hellwig <hch@infradead.org>,
Matthew Wilcox <willy@infradead.org>,
Michal Hocko <mhocko@kernel.org>,
Christopher Lameter <cl@linux.com>,
Jason Gunthorpe <jgg@ziepe.ca>,
Dan Williams <dan.j.williams@intel.com>,
linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>,
LKML <linux-kernel@vger.kernel.org>,
linux-rdma <linux-rdma@vger.kernel.org>,
linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH 4/6] mm: introduce page->dma_pinned_flags, _count
Date: Tue, 6 Nov 2018 13:47:15 +1100 [thread overview]
Message-ID: <20181106024715.GU6311@dastard> (raw)
In-Reply-To: <f5ad7210-05e0-3dc4-02df-01ce5346e198@nvidia.com>
On Mon, Nov 05, 2018 at 04:26:04PM -0800, John Hubbard wrote:
> On 11/5/18 1:54 AM, Jan Kara wrote:
> > Hmm, have you tried larger buffer sizes? Because synchronous 8k IO isn't
> > going to max-out NVME iops by far. Can I suggest you install fio [1] (it
> > has the advantage that it is pretty much standard for a test like this so
> > everyone knows what the test does from a glimpse) and run with it something
> > like the following workfile:
> >
> > [reader]
> > direct=1
> > ioengine=libaio
> > blocksize=4096
> > size=1g
> > numjobs=1
> > rw=read
> > iodepth=64
> >
> > And see how the numbers with and without your patches compare?
> >
> > Honza
> >
> > [1] https://github.com/axboe/fio
>
> That program is *very* good to have. Whew. Anyway, it looks like read bandwidth
> is approximately 74 MiB/s with my patch (it varies a bit, run to run),
> as compared to around 85 without the patch, so still showing about a 20%
> performance degradation, assuming I'm reading this correctly.
>
> Raw data follows, using the fio options you listed above:
>
> Baseline (without my patch):
> ----------------------------
....
> lat (usec): min=179, max=14003, avg=2913.65, stdev=1241.75
> clat percentiles (usec):
> | 1.00th=[ 2311], 5.00th=[ 2343], 10.00th=[ 2343], 20.00th=[ 2343],
> | 30.00th=[ 2343], 40.00th=[ 2376], 50.00th=[ 2376], 60.00th=[ 2376],
> | 70.00th=[ 2409], 80.00th=[ 2933], 90.00th=[ 4359], 95.00th=[ 5276],
> | 99.00th=[ 8291], 99.50th=[ 9110], 99.90th=[10945], 99.95th=[11469],
> | 99.99th=[12256]
.....
> Modified (with my patch):
> ----------------------------
.....
> lat (usec): min=81, max=15766, avg=3496.57, stdev=1450.21
> clat percentiles (usec):
> | 1.00th=[ 2835], 5.00th=[ 2835], 10.00th=[ 2835], 20.00th=[ 2868],
> | 30.00th=[ 2868], 40.00th=[ 2868], 50.00th=[ 2868], 60.00th=[ 2900],
> | 70.00th=[ 2933], 80.00th=[ 3425], 90.00th=[ 5080], 95.00th=[ 6259],
> | 99.00th=[10159], 99.50th=[11076], 99.90th=[12649], 99.95th=[13435],
> | 99.99th=[14484]
So it's adding at least 500us of completion latency to every IO?
I'd argue that the IO latency impact is far worse than the a 20%
throughput drop.
i.e. You can make up for throughput drops by running a deeper
queue/more dispatch threads, but you can't reduce IO latency at
all...
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
next prev parent reply other threads:[~2018-11-06 2:47 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-10-12 6:00 [PATCH 0/6] RFC: gup+dma: tracking dma-pinned pages john.hubbard
2018-10-12 6:00 ` [PATCH 1/6] mm: get_user_pages: consolidate error handling john.hubbard
2018-10-12 6:30 ` Balbir Singh
2018-10-12 22:45 ` John Hubbard
2018-10-12 6:00 ` [PATCH 2/6] mm: introduce put_user_page*(), placeholder versions john.hubbard
2018-10-12 7:35 ` Balbir Singh
2018-10-12 22:31 ` John Hubbard
2018-10-12 6:00 ` [PATCH 3/6] infiniband/mm: convert put_page() to put_user_page*() john.hubbard
2018-10-12 6:00 ` [PATCH 4/6] mm: introduce page->dma_pinned_flags, _count john.hubbard
2018-10-12 10:56 ` Balbir Singh
2018-10-13 0:15 ` John Hubbard
2018-10-24 11:00 ` Balbir Singh
2018-11-02 23:27 ` John Hubbard
2018-10-13 3:55 ` Dave Chinner
2018-10-13 7:34 ` John Hubbard
2018-10-13 16:47 ` Christoph Hellwig
2018-10-13 21:19 ` John Hubbard
2018-11-05 7:10 ` John Hubbard
2018-11-05 9:54 ` Jan Kara
2018-11-06 0:26 ` John Hubbard
2018-11-06 2:47 ` Dave Chinner [this message]
2018-11-06 11:00 ` Jan Kara
2018-11-06 20:41 ` Dave Chinner
2018-11-07 6:36 ` John Hubbard
2018-10-13 23:01 ` Dave Chinner
2018-10-16 8:51 ` Jan Kara
2018-10-17 1:48 ` John Hubbard
2018-10-17 11:09 ` Michal Hocko
2018-10-18 0:03 ` John Hubbard
2018-10-19 8:11 ` Michal Hocko
2018-10-12 6:00 ` [PATCH 5/6] mm: introduce zone_gup_lock, for dma-pinned pages john.hubbard
2018-10-12 6:00 ` [PATCH 6/6] mm: track gup pages with page->dma_pinned_* fields john.hubbard
2018-10-12 11:07 ` Balbir Singh
2018-10-13 0:33 ` John Hubbard
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20181106024715.GU6311@dastard \
--to=david@fromorbit.com \
--cc=akpm@linux-foundation.org \
--cc=cl@linux.com \
--cc=dan.j.williams@intel.com \
--cc=hch@infradead.org \
--cc=jack@suse.cz \
--cc=jgg@ziepe.ca \
--cc=jhubbard@nvidia.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-rdma@vger.kernel.org \
--cc=mhocko@kernel.org \
--cc=willy@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).