linux-nvme.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: Chaitanya Kulkarni <Chaitanya.Kulkarni@wdc.com>
To: Max Gurtovoy <maxg@mellanox.com>,
	"sagi@grimberg.me" <sagi@grimberg.me>, "hch@lst.de" <hch@lst.de>
Cc: "linux-nvme@lists.infradead.org" <linux-nvme@lists.infradead.org>
Subject: Re: [RFC PATCH 1/2] nvmet: add bdev-ns polling support
Date: Thu, 30 Jan 2020 18:19:43 +0000	[thread overview]
Message-ID: <BYAPR04MB57492FCA245C3F19079CDD5186040@BYAPR04MB5749.namprd04.prod.outlook.com> (raw)
In-Reply-To: <03494a32-b31a-432b-870d-7347f5cc9596@mellanox.com>

On 01/23/2020 06:23 AM, Max Gurtovoy wrote:
> On 1/21/2020 9:22 PM, Chaitanya Kulkarni wrote:
>> >Please see my in-line comments.
>> >On 01/20/2020 04:52 AM, Max Gurtovoy wrote:
>>> >>On 12/10/2019 8:25 AM, Chaitanya Kulkarni wrote:
>>>> >>>This patch adds support for the bdev-ns polling. We also add a new
>>>> >>>file to keep polling code separate (io-poll.c). The newly added
>>>> >>>configfs attribute allows user to enable/disable polling.
>>>> >>>
>>>> >>>We only enable polling for the READ/WRITE operation based on support
>>>> >>>from bdev and and use_poll configfs attr.
>>>> >>>
>>>> >>>We configure polling per namespace. For each namespace we create
>>>> >>>polling threads. For general approach please have a look at the
>>>> >>>cover-letter of this patch-series.
>>> >>It would be great to get some numbers here to learn about the trade-off
>>> >>for this approach.
>>> >>
>>> >>
>> >Already posted these numbers with QEMU nvme-loop here :-
>> >https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.infradead.org%2Fpipermail%2Flinux-nvme%2F2020-January%2F028692.html&amp;data=02%7C01%7Cmaxg%40mellanox.com%7C9308e8badf9f41ec4faf08d79ea733ae%7Ca652971c7d2e4d9ba6a4d149256f461b%7C0%7C0%7C637152313279111655&amp;sdata=RM3%2FcVeRcp8hgyAUFIM%2B7wbZ0cw2xvhdsfcVp9FuPm4%3D&amp;reserved=0
>> >
>> >IOPS/BW:-
>> >
>> >Default:-
>> >read: IOPS=52.8k, BW=206MiB/s (216MB/s)(6188MiB/30001msec)
>> >read: IOPS=54.3k, BW=212MiB/s (223MB/s)(6369MiB/30001msec)
>> >read: IOPS=53.5k, BW=209MiB/s (219MB/s)(6274MiB/30001msec)
>> >Poll:-
>> >read: IOPS=68.4k, BW=267MiB/s (280MB/s)(8011MiB/30001msec)
>> >read: IOPS=69.3k, BW=271MiB/s (284MB/s)(8124MiB/30001msec)
>> >read: IOPS=69.4k, BW=271MiB/s (284MB/s)(8132MiB/30001msec)
>> >
>> >mpstat:-
>> >Default:-
>> >CPU %usr %nice |%sys| %iowait %irq %soft %steal %guest %gnice %idle
>> >---------------------------------------------------------
>> >all 1.18 0.00 |60.14| 0.00 0.00 0.00 0.08 0.00 0.00 38.60
>> >---------------------------------------------------------
>> >0 0.00 0.00 | 32.00 |0.00 0.00 0.00 0.00 0.00 0.00 68.00
>> >1 1.01 0.00 | 42.42 |0.00 0.00 0.00 0.00 0.00 0.00 56.57
>> >2 1.01 0.00 | 57.58 |0.00 0.00 0.00 0.00 0.00 0.00 41.41
>> >3 2.02 0.00 | 79.80 |0.00 0.00 0.00 0.00 0.00 0.00 18.18
>> >4 1.01 0.00 | 63.64 |0.00 0.00 0.00 0.00 0.00 0.00 35.35
>> >5 3.09 0.00 | 63.92 |0.00 0.00 0.00 0.00 0.00 0.00 32.99
>> >6 2.02 0.00 | 75.76 |0.00 0.00 0.00 0.00 0.00 0.00 22.22
>> >7 1.02 0.00 | 57.14 |0.00 0.00 0.00 0.00 0.00 0.00 41.84
>> >8 0.00 0.00 | 67.01 |0.00 0.00 0.00 0.00 0.00 0.00 32.99
>> >9 1.01 0.00 | 59.60 |0.00 0.00 0.00 0.00 0.00 0.00 39.39
>> >10 1.02 0.00| 62.24 |0.00 0.00 0.00 0.00 0.00 0.00 36.73
>> >11 1.02 0.00| 62.24 |0.00 0.00 0.00 0.00 0.00 0.00 36.73
>> >
>> >Poll:-
>> >CPU %usr %nice |%sys| %iowait %irq %soft %steal %guest %gnice %idle
>> >---------------------------------------------------------
>> >all 1.08 0.00 98.08 0.00 0.00 0.00 0.08 0.00 0.00 0.75
>> >---------------------------------------------------------
>> >0 2.00 0.00 | 97.00 |0.00 0.00 0.00 0.00 0.00 0.00 1.00
>> >1 0.00 0.00 | 97.00 |0.00 0.00 0.00 0.00 0.00 0.00 3.00
>> >2 1.00 0.00 | 99.00 |0.00 0.00 0.00 0.00 0.00 0.00 0.00
>> >3 1.00 0.00 | 99.00 |0.00 0.00 0.00 0.00 0.00 0.00 0.00
>> >4 1.01 0.00 | 98.99 |0.00 0.00 0.00 0.00 0.00 0.00 0.00
>> >5 1.00 0.00 | 99.00 |0.00 0.00 0.00 0.00 0.00 0.00 0.00
>> >6 0.99 0.00 | 99.01 |0.00 0.00 0.00 0.00 0.00 0.00 0.00
>> >7 1.00 0.00 | 97.00 |0.00 0.00 0.00 0.00 0.00 0.00 2.00
>> >8 1.00 0.00 | 99.00 |0.00 0.00 0.00 0.00 0.00 0.00 0.00
>> >9 1.00 0.00 | 99.00 |0.00 0.00 0.00 0.00 0.00 0.00 0.00
>> >10 1.98 0.00| 94.06 |0.00 0.00 0.00 0.00 0.00 0.00 3.96
>> >11 1.00 0.00| 99.00 |0.00 0.00 0.00 0.00 0.00 0.00 0.00
> with all the respect to nvmet-loop, it isn't a real use case, right ?
>
> Can you re-run it using RDMA and TCP ?
>
> what is the backing device for these tests ? NVMe/NULL_BLK ?
>
>
I don't think null_blk supports polling. This is NVMe QEMU Device.

Don't have the RDMA setup, I'll create one and report numbers.






_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

  reply	other threads:[~2020-01-30 18:20 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-12-10  6:25 [RFC PATCH 0/2] nvmet: add " Chaitanya Kulkarni
2019-12-10  6:25 ` [RFC PATCH 1/2] nvmet: add bdev-ns " Chaitanya Kulkarni
2020-01-20 12:52   ` Max Gurtovoy
2020-01-21 19:22     ` Chaitanya Kulkarni
2020-01-23 14:23       ` Max Gurtovoy
2020-01-30 18:19         ` Chaitanya Kulkarni [this message]
2019-12-10  6:25 ` [RFC PATCH 2/2] nvmet: add file-ns " Chaitanya Kulkarni
2019-12-12  1:01 ` [RFC PATCH 0/2] nvmet: add " Sagi Grimberg
2019-12-12  5:44   ` Chaitanya Kulkarni
2019-12-12 20:32     ` Sagi Grimberg
2020-01-20  5:13       ` Chaitanya Kulkarni
2020-01-20  4:48   ` Chaitanya Kulkarni

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=BYAPR04MB57492FCA245C3F19079CDD5186040@BYAPR04MB5749.namprd04.prod.outlook.com \
    --to=chaitanya.kulkarni@wdc.com \
    --cc=hch@lst.de \
    --cc=linux-nvme@lists.infradead.org \
    --cc=maxg@mellanox.com \
    --cc=sagi@grimberg.me \
    --subject='Re: [RFC PATCH 1/2] nvmet: add bdev-ns polling support' \
    /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

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).