qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Dmitry Zhurikhin <zhur@ispras.ru>
To: Aurelien Jarno <aurelien@aurel32.net>
Cc: mj.mccormack@samsung.com, qemu-devel@nongnu.org,
	Kirill Batuzov <batuzovk@ispras.ru>
Subject: Re: [Qemu-devel] [PATCH 0/6] Implement constant folding and copy propagation in TCG
Date: Sat, 21 May 2011 16:47:37 +0400	[thread overview]
Message-ID: <4DD7B469.1040802@ispras.ru> (raw)
In-Reply-To: <20110520193508.GB27170@hall.aurel32.net>

On 05/20/2011 11:35 PM, Aurelien Jarno wrote:
> 
> On Fri, May 20, 2011 at 04:39:27PM +0400, Kirill Batuzov wrote:
>> This series implements some basic machine-independent optimizations.  They
>> simplify code and allow liveness analysis do it's work better.
>>
>> Suppose we have following ARM code:
>>
>>  movw    r12, #0xb6db
>>  movt    r12, #0xdb6d
>>
>> In TCG before optimizations we'll have:
>>
>>  movi_i32 tmp8,$0xb6db
>>  mov_i32 r12,tmp8
>>  mov_i32 tmp8,r12
>>  ext16u_i32 tmp8,tmp8
>>  movi_i32 tmp9,$0xdb6d0000
>>  or_i32 tmp8,tmp8,tmp9
>>  mov_i32 r12,tmp8
>>
>> And after optimizations we'll have this:
>>
>>  movi_i32 r12,$0xdb6db6db
>>
>> Here are performance evaluation results on SPEC CPU2000 integer tests in
>> user-mode emulation on x86_64 host.  There were 5 runs of each test on
>> reference data set.  The tables below show runtime in seconds for all these
>> runs.
> 
> How are the tests done? Are they done with linux-user, or running the
> executables in qemu-system-xxx?
They were run in user mode on a dedicated machine not doing anything
else.  We found system emulation to be too volatile for measuring
anything.  Anyway even with user mode and on a dedicated machine there
are some weird performance jumps we can't explain but overall SPEC seems
stable enough to notice influence of such changes in code generation.

> 
>> ...

	Dmitry

  reply	other threads:[~2011-05-21 12:47 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-05-20 12:39 [Qemu-devel] [PATCH 0/6] Implement constant folding and copy propagation in TCG Kirill Batuzov
2011-05-20 12:39 ` [Qemu-devel] [PATCH 1/6] Add TCG optimizations stub Kirill Batuzov
2011-05-20 18:12   ` Richard Henderson
2011-05-20 18:33     ` Richard Henderson
2011-05-20 12:39 ` [Qemu-devel] [PATCH 2/6] Add copy and constant propagation Kirill Batuzov
2011-05-20 18:22   ` Richard Henderson
2011-05-20 18:46     ` Paolo Bonzini
2011-05-20 19:41   ` Aurelien Jarno
2011-05-20 12:39 ` [Qemu-devel] [PATCH 3/6] Do constant folding for basic arithmetic operations Kirill Batuzov
2011-05-20 12:39 ` [Qemu-devel] [PATCH 4/6] Do constant folding for boolean operations Kirill Batuzov
2011-05-20 18:45   ` Richard Henderson
2011-05-20 12:39 ` [Qemu-devel] [PATCH 5/6] Do constant folding for shift operations Kirill Batuzov
2011-05-20 18:37   ` Richard Henderson
2011-05-26 12:36     ` Kirill Batuzov
2011-05-26 13:56       ` Richard Henderson
2011-05-26 19:14         ` Blue Swirl
2011-05-26 20:10           ` Richard Henderson
2011-05-26 20:25             ` Blue Swirl
2011-05-26 21:14               ` Richard Henderson
2011-05-27 15:41                 ` Jamie Lokier
2011-05-27 17:07                 ` Blue Swirl
2011-05-27 19:54                   ` Richard Henderson
2011-05-27  7:09           ` Paolo Bonzini
2011-05-20 12:39 ` [Qemu-devel] [PATCH 6/6] Do constant folding for unary operations Kirill Batuzov
2011-05-20 18:39   ` Richard Henderson
2011-05-20 17:50 ` [Qemu-devel] [PATCH 0/6] Implement constant folding and copy propagation in TCG Richard Henderson
2011-05-20 19:37   ` Aurelien Jarno
2011-05-20 23:31     ` Andreas Färber
2011-05-21  9:37       ` Laurent Desnogues
2011-05-21 10:46         ` Aurelien Jarno
2011-05-21 17:53           ` Kirill Batuzov
2011-05-20 19:35 ` Aurelien Jarno
2011-05-21 12:47   ` Dmitry Zhurikhin [this message]
2011-05-21 12:48   ` Aurelien Jarno

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=4DD7B469.1040802@ispras.ru \
    --to=zhur@ispras.ru \
    --cc=aurelien@aurel32.net \
    --cc=batuzovk@ispras.ru \
    --cc=mj.mccormack@samsung.com \
    --cc=qemu-devel@nongnu.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).