Linux MIPS Architecture development
 help / color / mirror / Atom feed
From: cgd@broadcom.com
To: hjl@lucon.org
Cc: "Justin Carlson" <justinca@ri.cmu.edu>,
	"Daniel Jacobowitz" <dan@debian.org>,
	"Maciej W. Rozycki" <macro@ds2.pg.gda.pl>,
	"Hiroyuki Machida" <machida@sm.sony.co.jp>,
	libc-alpha@sources.redhat.com, linux-mips@oss.sgi.com,
	gcc@gcc.gnu.org
Subject: Re: PATCH: Fix ll/sc for mips (take 3)
Date: 03 Feb 2002 15:29:28 -0800	[thread overview]
Message-ID: <yov5ofj65elj.fsf@broadcom.com> (raw)
In-Reply-To: hjl@lucon.org's message of "Sat, 2 Feb 2002 20:04:10 +0000 (UTC)"

[ Hi Justin, enjoying winter in pgh?  8-]

At Sat, 2 Feb 2002 20:04:10 +0000 (UTC), "H . J . Lu" wrote:
> Does everyone agree with this? If yes, I can make a patch not to use
> branch likely. But on the other hand, "gcc -mips2" will generate code
> using branch likely. If branch likely doesn't buy you anything, 
> shouldn't we change gcc not to generate branch likely instructions?

Branch-likely instructions probably _do_ buy you something (at least,
slightly less code size) on some CPUs, probably even some CPUs which
are still being produced.

FYI, somebody, i think it was Mike Stump, submitted a patch to add a
flag to disable branch-likely instructions N months ago.  It was
discussed a little bit, not AFAIK nothing was ever done with it.


To quote the MIPS32 instruction set document from the MIPS web site:

> Programming Notes:
> 
> [ ... ]
>
> Software is strongly encouraged to avoid the use of the Branch
> Likely instructions, as they will be removed from a future revision
> of the MIPS Architecture.
>
> Some implementations always predict the branch will be taken, so
> there is a significant penalty if the branch is not taken. Software
> should only use this instruction when there is a very high
> probability (98% or more) that the branch will be taken. If the
> branch is not likely to be taken or if the probability of a taken
> branch is unknown, software is encouraged to use the BGEZAL
> instruction instead.

If you have a sufficiently high static likelyhood of branch-taken, it
may well be worthwhile to use branch-likely even on MIPS32/MIPS64 if
it'll save you code space, or if you can determine that not clogging
up your branch history / predictor buffers makes your code run
faster.  However, it should definitely not be the default.



Anyway, from where I sit, this should be resolved like:

* -mips1 - -mips5 should generate them by default (for
  strongly-predicted branches), -mips32 & -mips64 should not.  (modify
  that for your favorite flag names, -march or whatever.  8-)

* There should be flags to override the defaults either way, so you
  could say -mips2 -mno-branch-likely, or -mips32 -mbranch-likely.


cgd

  parent reply	other threads:[~2002-02-04  0:29 UTC|newest]

Thread overview: 75+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-01-31 20:35 PATCH: Fix ll/sc for mips H . J . Lu
2002-01-31 22:17 ` Maciej W. Rozycki
2002-01-31 22:41   ` H . J . Lu
2002-02-01  3:35     ` Hiroyuki Machida
2002-02-01  4:02       ` [libc-alpha] " Kaz Kylheku
2002-02-01  4:02         ` Kaz Kylheku
2002-02-01  4:59         ` Hiroyuki Machida
2002-02-01  7:00           ` H . J . Lu
2002-02-01 11:12             ` Hiroyuki Machida
2002-02-01 10:49           ` Andreas Schwab
2002-02-01 11:23             ` Geoff Keating
2002-02-01  7:10         ` H . J . Lu
2002-02-01  7:17       ` H . J . Lu
2002-02-01 11:45         ` Maciej W. Rozycki
2002-02-01 18:29           ` PATCH: Fix ll/sc for mips (take 3) H . J . Lu
2002-02-01 23:01             ` Daniel Jacobowitz
2002-02-01 23:15               ` H . J . Lu
2002-02-02  2:37                 ` Hiroyuki Machida
2002-02-04  9:32                   ` Dominic Sweetman
2002-02-05  6:16                     ` Jay Carlson
2002-02-05  8:28                       ` Ralf Baechle
2002-02-05 15:10                         ` Jay Carlson
2002-02-05 16:06                           ` Jay Carlson
2002-02-02  3:26                 ` Daniel Jacobowitz
2002-02-02 18:53                   ` Justin Carlson
2002-02-02 20:03                     ` H . J . Lu
2002-02-02 20:49                       ` Hartvig Ekner
2002-02-02 20:49                         ` Hartvig Ekner
2002-02-03  5:47                       ` Justin Carlson
2002-02-04 19:17                         ` Paul Koning
     [not found]                       ` <mailpost.1012680250.7159@news-sj1-1>
2002-02-03 23:29                         ` cgd [this message]
2002-02-04  6:07                           ` Ralf Baechle
2002-02-04  9:46                           ` Dominic Sweetman
2002-02-04 16:31                             ` H . J . Lu
2002-02-04 16:46                               ` Dominic Sweetman
2002-02-05  1:28                                 ` H . J . Lu
2002-02-05  2:58                                   ` Daniel Jacobowitz
2002-02-05  4:42                                     ` H . J . Lu
2002-02-05  4:47                                       ` Daniel Jacobowitz
2002-02-05  5:30                                         ` Justin Carlson
2002-02-05  8:39                                           ` Hartvig Ekner
2002-02-05  8:39                                             ` Hartvig Ekner
2002-02-05 11:37                                             ` Maciej W. Rozycki
2002-02-05 12:12                                             ` Ralf Baechle
2002-02-05 12:31                                               ` Maciej W. Rozycki
2002-02-05 12:38                                                 ` Hartvig Ekner
2002-02-05 12:38                                                   ` Hartvig Ekner
2002-02-05 13:28                                                   ` Maciej W. Rozycki
2002-02-05 19:28                                                     ` Hartvig Ekner
2002-02-05 19:28                                                       ` Hartvig Ekner
2002-02-05 18:59                                                   ` Ralf Baechle
2002-02-05 19:30                                     ` H . J . Lu
2002-02-05 21:54                                       ` H . J . Lu
2002-02-06 10:32                                         ` Ralf Baechle
2002-02-06 20:45                                           ` Why does -g turn off filling the delat slot? H . J . Lu
2002-02-06 21:00                                             ` PATCH: Modify the mips gas behavior for -g -O H . J . Lu
2002-02-06 21:16                                               ` Eric Christopher
2002-02-06 21:40                                                 ` Ian Lance Taylor
2002-02-06 21:46                                                   ` Eric Christopher
2002-02-06 22:00                                                   ` PATCH: Define SUBTARGET_ASM_DEBUGGING_SPEC for Linux/mips H . J . Lu
2002-02-07  8:24                                                     ` Eric Christopher
2002-02-06 11:37                                         ` PATCH: Fix ll/sc for mips (take 3) Maciej W. Rozycki
2002-02-04 17:44                             ` cgd
2002-02-04  6:46                     ` Ralf Baechle
2002-02-04  7:01         ` PATCH: Fix ll/sc for mips Ralf Baechle
2002-02-04 14:54           ` Maciej W. Rozycki
2002-02-01 11:50       ` Maciej W. Rozycki
2002-02-01 17:40         ` H . J . Lu
2002-02-01 21:41           ` Maciej W. Rozycki
2002-02-01 22:47             ` H . J . Lu
2002-02-02 11:06               ` Maciej W. Rozycki
2002-02-03  2:29               ` Ulrich Drepper
2002-02-03  2:29                 ` Ulrich Drepper
2002-01-31 23:33   ` [libc-alpha] " Kaz Kylheku
2002-01-31 23:33     ` Kaz Kylheku

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=yov5ofj65elj.fsf@broadcom.com \
    --to=cgd@broadcom.com \
    --cc=dan@debian.org \
    --cc=gcc@gcc.gnu.org \
    --cc=hjl@lucon.org \
    --cc=justinca@ri.cmu.edu \
    --cc=libc-alpha@sources.redhat.com \
    --cc=linux-mips@oss.sgi.com \
    --cc=machida@sm.sony.co.jp \
    --cc=macro@ds2.pg.gda.pl \
    /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