From: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
To: Minchan Kim <minchan@kernel.org>
Cc: Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>,
Sergey Senozhatsky <sergey.senozhatsky@gmail.com>,
Andrew Morton <akpm@linux-foundation.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
Linux-MM <linux-mm@kvack.org>, Nitin Gupta <ngupta@vflare.org>,
Jerome Marchand <jmarchan@redhat.com>,
Ganesh Mahendran <opensource.ganesh@gmail.com>
Subject: Re: [PATCH v1 2/2] zram: remove init_lock in zram_make_request
Date: Sat, 31 Jan 2015 20:31:58 +0900 [thread overview]
Message-ID: <20150131113158.GB2299@swordfish> (raw)
In-Reply-To: <20150130144145.GA2840@blaptop>
Hello Minchan,
excellent analysis!
On (01/30/15 23:41), Minchan Kim wrote:
> Yes, __srcu_read_lock is a little bit heavier but the number of instruction
> are not too much difference to make difference 10%. A culprit is
> __cond_resched but I don't think, either because our test was CPU intensive
> soS I don't think schedule latency affects total bandwidth.
>
> More cuprit is your data pattern.
> It seems you didn't use scramble_buffers=0, zero_buffers in fio so that
> fio fills random data pattern so zram bandwidth could be different by
> compression/decompression ratio.
Completely agree.
Shame on me. gotten so used to iozone (iozone uses same data pattern 0xA5,
this is +Z option what for), so I didn't even think about data pattern
in fio. sorry.
> 1) randread
> srcu is worse as 0.63% but the difference is really marginal.
>
> 2) randwrite
> srcu is better as 1.24% is better.
>
> 3) randrw
> srcu is better as 2.3%
hm, interesting. I'll re-check.
> Okay, if you concerns on the data still, how about this?
I'm not so upset to lose 0.6234187%. my concerns were about iozone's
10% different (which looks a bit worse).
I'll review your patch. Thanks for your effort.
> >
> > by "data pattern" you mean usage scenario? well, I usually use zram for
> > `make -jX', where X=[4..N]. so N concurrent read-write ops scenario.
>
> What I meant is what data fills I/O buffer, which is really important
> to evaluate zram because the compression/decompression speeds relys on it.
>
I see. I never test it with `make' anyway, only iozone +Z.
-ss
next prev parent reply other threads:[~2015-01-31 11:31 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <1422432945-6764-1-git-send-email-minchan@kernel.org>
2015-01-28 14:19 ` [PATCH 1/2] zram: free meta table in zram_meta_free Sergey Senozhatsky
2015-01-28 23:17 ` Minchan Kim
2015-01-29 1:49 ` Ganesh Mahendran
[not found] ` <1422432945-6764-2-git-send-email-minchan@kernel.org>
2015-01-28 14:56 ` [PATCH v1 2/2] zram: remove init_lock in zram_make_request Sergey Senozhatsky
2015-01-28 15:04 ` Sergey Senozhatsky
2015-01-28 23:33 ` Minchan Kim
[not found] ` <CAHqPoqKZFDSjO1pL+ixYe_m_L0nGNcu04qSNp-jd1fUixKtHnw@mail.gmail.com>
2015-01-29 2:01 ` Minchan Kim
2015-01-29 2:22 ` Sergey Senozhatsky
2015-01-29 5:28 ` Minchan Kim
2015-01-29 6:06 ` Sergey Senozhatsky
2015-01-29 6:35 ` Minchan Kim
2015-01-29 7:08 ` Sergey Senozhatsky
2015-01-30 14:41 ` Minchan Kim
2015-01-31 11:31 ` Sergey Senozhatsky [this message]
2015-02-01 14:50 ` Sergey Senozhatsky
2015-02-01 15:04 ` Sergey Senozhatsky
2015-02-02 1:43 ` Minchan Kim
2015-02-02 1:59 ` Sergey Senozhatsky
2015-02-02 2:45 ` Minchan Kim
2015-02-02 3:47 ` Sergey Senozhatsky
2015-02-02 1:30 ` Minchan Kim
2015-02-02 1:48 ` Sergey Senozhatsky
2015-02-02 2:44 ` Minchan Kim
2015-02-02 4:01 ` Sergey Senozhatsky
2015-02-02 4:28 ` Minchan Kim
2015-02-02 5:09 ` Sergey Senozhatsky
2015-02-02 5:18 ` Minchan Kim
2015-02-02 5:28 ` Sergey Senozhatsky
2015-02-02 5:10 ` Minchan Kim
2015-01-30 0:20 ` Sergey Senozhatsky
2015-01-29 13:48 ` Ganesh Mahendran
2015-01-29 15:12 ` Sergey Senozhatsky
2015-01-30 7:52 ` Ganesh Mahendran
2015-01-30 8:08 ` Sergey Senozhatsky
2015-01-31 8:50 ` Ganesh Mahendran
2015-01-31 11:07 ` Sergey Senozhatsky
2015-01-31 12:59 ` Ganesh Mahendran
2015-02-02 3:41 Minchan Kim
2015-02-02 5:59 ` Sergey Senozhatsky
2015-02-02 6:18 ` Minchan Kim
2015-02-02 7:06 ` Sergey Senozhatsky
2015-02-03 1:54 ` Sergey Senozhatsky
2015-02-03 3:02 ` Minchan Kim
2015-02-03 3:56 ` Sergey Senozhatsky
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=20150131113158.GB2299@swordfish \
--to=sergey.senozhatsky@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=jmarchan@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=minchan@kernel.org \
--cc=ngupta@vflare.org \
--cc=opensource.ganesh@gmail.com \
--cc=sergey.senozhatsky.work@gmail.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;
as well as URLs for NNTP newsgroup(s).