From: Jens Axboe <axboe@kernel.dk>
To: Andrey Kuzmin <andrey.v.kuzmin@gmail.com>,
Jeff Furlong <jeff.furlong@hgst.com>
Cc: "fio@vger.kernel.org" <fio@vger.kernel.org>
Subject: Re: Specify range and distribution of accesses
Date: Mon, 7 Mar 2016 19:41:06 -0700 [thread overview]
Message-ID: <56DE3BC2.6040206@kernel.dk> (raw)
In-Reply-To: <CANvN+e=jaYQZNS7KRHALDeEF6bwJ0uqDf7ra+qoCwy5OYXMzwA@mail.gmail.com>
On 03/07/2016 02:02 PM, Andrey Kuzmin wrote:
> On Mon, Mar 7, 2016 at 11:46 PM, Jeff Furlong <jeff.furlong@hgst.com> wrote:
>> Thanks for the suggestions and patches. Using the latest fio version, the JESD219 workload is possible:
>
> Nice.
>
>>
>> # fio -version
>> fio-2.6-27-gd283
>>
>> # fio --name=JESD219 --ioengine=libaio --direct=1 --rw=randrw --norandommap --randrepeat=0 --rwmixread=40 --rwmixwrite=60 --iodepth=256 --size=100% --numjobs=4 --bssplit=512/4:1024/1:1536/1:2048/1:2560/1:3072/1:3584/1:4k/67:8k/10:16k/7:32k/3:64k/3 --random_distribution=zoned:50/5:30/15:20/80 --overwrite=1 --filename=/dev/nvme0n1 --group_reporting --runtime=5m --time_based --output=JESD219
>>
>> A quick statistical analysis of the results shows:
>>
>> Found 20380582 IOs
>>
>> Found 39.9903152913% reads
>> Found 60.0096847087% writes
>>
>> Found 4.00492979052% 512
>> Found 1.00495658073% 1024
>> Found 1.00079575745% 1536
>> Found 1.00046701316% 2048
>> Found 0.998764412125% 2560
>> Found 0.998043137335% 3072
>> Found 0.999520033334% 3584
>> Found 67.0145778958% 4096
>> Found 9.98662844859% 8192
>> Found 6.99898560306% 16384
>> Found 2.99961993235% 32768
>> Found 2.99271139558% 65536
>>
>> Found 49.9895734086% 0-5%
>> Found 30.0126463513% 5-20%%
>> Found 19.99778024% 20-100%
>>
>
> It hardly matters, but is still somewhat surprising to see that both
> bs and zone split percentage are accurate only up to 5x10^-3.
So I did the math on these, an fio is on (or really close) to the
expected deviation for random outcomes.
The first two of the block sizes (didn't check more):
> Found 4.00492979052% 512
3.99387% to 4.00612% would be in the range.
> Found 1.00495658073% 1024
0.9938778% to 1.00612% would be in the range.
For the access location:
> Found 49.9895734086% 0-5%
49.98955% to 50.01044% would be in the range.
> Found 30.0126463513% 5-20%%
29.98955% to 30.01044% would be in the range.
> Found 19.99778024% 20-100%
19.98955% to 20.01044% would be in the range.
So while the deviations from the specified does seem larger than
intuition would lead you to believe, it's actually really close. I think
we can call this one closed.
--
Jens Axboe
next prev parent reply other threads:[~2016-03-08 2:41 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-02-26 20:53 Specify range and distribution of accesses Jeff Furlong
2016-02-27 9:14 ` Andrey Kuzmin
2016-03-03 16:18 ` Jens Axboe
2016-03-03 20:04 ` Jens Axboe
2016-03-07 20:46 ` Jeff Furlong
2016-03-07 21:02 ` Andrey Kuzmin
2016-03-07 21:08 ` Jens Axboe
2016-03-08 2:41 ` Jens Axboe [this message]
2016-03-07 21:08 ` Jens Axboe
2016-03-07 21:19 ` Elliott, Robert (Persistent Memory)
2016-03-07 21:45 ` Jens Axboe
2016-03-12 2:07 ` Vladislav Bolkhovitin
2016-03-03 16:16 ` Jens Axboe
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=56DE3BC2.6040206@kernel.dk \
--to=axboe@kernel.dk \
--cc=andrey.v.kuzmin@gmail.com \
--cc=fio@vger.kernel.org \
--cc=jeff.furlong@hgst.com \
/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