public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
* Re: Some thoughts about steerring commitee work
@ 2007-06-16  2:03 J.C. Pizarro
  2007-06-16  2:43 ` Ian Lance Taylor
                   ` (2 more replies)
  0 siblings, 3 replies; 58+ messages in thread
From: J.C. Pizarro @ 2007-06-16  2:03 UTC (permalink / raw)
  To: gcc

Please, to see

1. "The LLVM Compiler System" by Chris Lattner
http://llvm.org/pubs/2007-03-12-BossaLLVMIntro.html
http://llvm.org/pubs/2007-03-12-BossaLLVMIntro.pdf

2. "Vector LLVA: A Virtual Vector Instruction Set for Media Processing"
by Bocchino and Vikram
http://llvm.org/pubs/2006-06-15-VEE-VectorLLVA.html
http://llvm.org/pubs/2006-06-15-VEE-VectorLLVA.pdf

The LLVM follows the K.I.S.S. principle, GCC doesn't.

The optimizing stages of GCC's backend are big, fragmented and complex.

I think that the GCC's commitee goes to the wrong direction.

^ permalink raw reply	[flat|nested] 58+ messages in thread
* Re: Some thoughts about steerring commitee work
@ 2007-06-25 13:07 Mircea Namolaru
  0 siblings, 0 replies; 58+ messages in thread
From: Mircea Namolaru @ 2007-06-25 13:07 UTC (permalink / raw)
  To: vmakarov; +Cc: gcc

>  As for C++, I think we need more OO language specific 
> optimizations. I don't know what the status of 
> devirtualizion which was reported on the previous 
> summit.

Sorry for the late replay.

The devirtualization is on hold. Currently GCC is lacking the necessary 
infrastructure needed by C++ interprocedural optimizations, like 
devirtualization. 

The appropriate infrastructure will be provided by LTO. I intend to 
incorporate the devirtualization in the LTO framework.

Mircea

infrastructure. 

^ permalink raw reply	[flat|nested] 58+ messages in thread
* Re: Some thoughts about steerring commitee work
@ 2007-06-18 17:16 Devang Patel
  0 siblings, 0 replies; 58+ messages in thread
From: Devang Patel @ 2007-06-18 17:16 UTC (permalink / raw)
  To: Daniel Berlin; +Cc: DORIT, gcc

>> I can hand you more than the testcases i've given so far.  There is
>> tons of code out there that would  benefit from straight line
>> vectorization.

I'm interested in these test cases. Thanks!

>> In fact, we have some that gets written in loop form
>> right now just so it gets vectorized!

May be loop materialization is useful in such situation ?

-
Devang

^ permalink raw reply	[flat|nested] 58+ messages in thread
* Re: Some thoughts about steerring commitee work
@ 2007-06-17 13:59 Ira Rosen
  2007-06-17 16:11 ` Daniel Berlin
  0 siblings, 1 reply; 58+ messages in thread
From: Ira Rosen @ 2007-06-17 13:59 UTC (permalink / raw)
  To: Daniel Berlin
  Cc: Dorit Nuzman, Vladimir N. Makarov, Eric Botcazou, gcc, H. J. Lu,
	Ian Lance Taylor


"Daniel Berlin" <dberlin@dberlin.org> wrote on 16/06/2007:

> On 6/16/07, Dorit Nuzman <DORIT@il.ibm.com> wrote:
>
> > Do you have specific examples where SLP helps performance out of loops?
>
> hash calculations.
>
> For md5, you can get a 2x performance improvement by straight-line
> vectorizing it
> sha1 is about 2-2.5x
>
> (This assumes you do good pack/unpack placement using something like
> lazy code motion)
>
> See, for example, http://arctic.org/~dean/crypto/sha1.html
>
> (The page is out of date, the technique they explain where they are
> doing straight line computation of the hash in parallel, is exactly
> what SLP would provide out of loops)

I looked at the above page (and also at MD5 and SHA1 implementations). I
found only computations inside loops.
Could you please explain what exactly you refer to as SLP out of loops in
this benchmark?

Thanks,
Ira

^ permalink raw reply	[flat|nested] 58+ messages in thread
* Diego Novillo appointed middle-end maintainer and non-algorithmic GWP
@ 2007-06-14 10:37 David Edelsohn
  2007-06-15 12:49 ` Some thoughts about steerring commitee work Vladimir N. Makarov
  0 siblings, 1 reply; 58+ messages in thread
From: David Edelsohn @ 2007-06-14 10:37 UTC (permalink / raw)
  To: dnovillo, gcc

	I am pleased to announce that the GCC Steering Committee has
promoted Diego Novillo to middle-end maintainer and appointed him
non-algorithmic Blanket/Global Write Privileges maintainer.

	Please join me in congratulating Diego on his
new role.  Please update your listings in the MAINTAINERS file.

Happy hacking!
David

^ permalink raw reply	[flat|nested] 58+ messages in thread

end of thread, other threads:[~2007-09-09 20:24 UTC | newest]

Thread overview: 58+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-06-16  2:03 Some thoughts about steerring commitee work J.C. Pizarro
2007-06-16  2:43 ` Ian Lance Taylor
2007-06-16  3:32 ` Richard Kenner
2007-06-16 19:08 ` Joe Buck
  -- strict thread matches above, loose matches on Subject: below --
2007-06-25 13:07 Mircea Namolaru
2007-06-18 17:16 Devang Patel
2007-06-17 13:59 Ira Rosen
2007-06-17 16:11 ` Daniel Berlin
2007-06-18 10:59   ` Dorit Nuzman
2007-06-18 12:01     ` Daniel Berlin
2007-06-18 12:53       ` Dorit Nuzman
2007-06-18 17:57         ` Sebastian Pop
2007-06-18 18:10           ` Daniel Berlin
2007-06-18 20:54     ` Jagasia, Harsha
2007-06-14 10:37 Diego Novillo appointed middle-end maintainer and non-algorithmic GWP David Edelsohn
2007-06-15 12:49 ` Some thoughts about steerring commitee work Vladimir N. Makarov
2007-06-15 13:05   ` Richard Kenner
2007-06-15 13:58     ` Vladimir N. Makarov
2007-06-15 17:07   ` Joe Buck
2007-06-15 17:48     ` Vladimir N. Makarov
2007-06-15 17:56       ` Joe Buck
2007-09-09 20:24       ` Gerald Pfeifer
2007-06-15 18:43   ` Ian Lance Taylor
2007-06-15 19:28     ` Vladimir N. Makarov
2007-06-15 19:41       ` Ian Lance Taylor
2007-06-15 20:55         ` Vladimir N. Makarov
2007-06-15 21:08           ` Eric Botcazou
2007-06-15 21:13             ` Daniel Berlin
2007-06-15 21:43               ` Eric Botcazou
2007-06-15 22:16                 ` Vladimir N. Makarov
2007-06-15 23:04               ` Richard Kenner
2007-06-15 21:36             ` Vladimir N. Makarov
2007-06-16  0:51               ` Tobias Burnus
2007-06-17 14:33               ` Paolo Bonzini
2007-06-16  1:43             ` Ian Lance Taylor
2007-06-16  2:02               ` H. J. Lu
2007-06-16  2:19                 ` Andrew Pinski
2007-06-16 13:25                   ` H. J. Lu
2007-06-16  3:33                 ` Vladimir N. Makarov
2007-06-16 15:04                   ` H. J. Lu
2007-06-16 16:13                     ` Dorit Nuzman
2007-06-17  0:48                       ` H. J. Lu
2007-06-17  1:16                         ` Tim Prince
2007-06-17  8:30                           ` Dorit Nuzman
2007-06-16 15:53                   ` Dorit Nuzman
2007-06-16 16:46                     ` Daniel Berlin
2007-06-16 15:29                 ` Dorit Nuzman
2007-06-16 23:59                 ` Ryan Hill
2007-06-17 10:44                   ` Dorit Nuzman
2007-06-17 14:52                     ` Ryan Hill
2007-06-18 10:47                       ` Dorit Nuzman
2007-06-16  3:08               ` Vladimir N. Makarov
2007-06-16  3:36                 ` Richard Kenner
2007-06-16  1:54           ` Ian Lance Taylor
2007-06-16  3:28             ` Vladimir N. Makarov
2007-06-16  3:45               ` Richard Kenner
2007-06-16 15:20                 ` Joel Sherrill
2007-06-18 23:12                   ` Mark Mitchell
2007-06-19 21:01                     ` Toon Moene

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 read-only IMAP folder(s) and NNTP newsgroup(s).