all messages for Emacs-related lists mirrored at yhetil.org
 help / color / mirror / code / Atom feed
From: Tom Tromey <tom@tromey.com>
To: "Clément Pit-Claudel" <cpitclaudel@gmail.com>
Cc: Tom Tromey <tom@tromey.com>, emacs-devel@gnu.org
Subject: Re: Emacs Lisp JIT Compiler
Date: Wed, 15 Aug 2018 18:32:35 -0600	[thread overview]
Message-ID: <877ekri21o.fsf@tromey.com> (raw)
In-Reply-To: <17e1d5ea-d8f7-9c94-979b-e89d119fe72c@gmail.com> ("Clément Pit-Claudel"'s message of "Tue, 14 Aug 2018 20:21:49 -0400")

>>>>> "Clément" == Clément Pit-Claudel <cpitclaudel@gmail.com> writes:

Clément> I compiled and installed libjit, and things seem to build fine here.

Thank you for trying it.

Clément> Benchmarking a simple loop yields an impressive 4x speedup (from 1
Clément> second down to 0.25s).  However, benchmarking a larger application
Clément> seems to yield a 30% slowdown (from 5.5s to 7.5s).  I haven't
Clément> investigated why yet.  Have you seen similar variations?

I have not really done that much benchmarking.
However, please send me your test case (or tell me what it is) and I
will take a look.

I'd expect the JIT to be at its best for some kinds of integer code and
things like walking lists -- it knows how to emit some simple things
inline.

One reason it could be slower is if the compiler is compiling many
functions that don't often run.  Many JITs have some heuristic about
when to compile, this could be done here as well.

Tom



  reply	other threads:[~2018-08-16  0:32 UTC|newest]

Thread overview: 60+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-08-13  4:01 Emacs Lisp JIT Compiler Tom Tromey
2018-08-13  5:37 ` Paul Eggert
2018-08-13 15:15   ` Tom Tromey
2018-08-14  0:16   ` Tom Tromey
2018-08-14 20:11     ` Daniel Colascione
2018-08-14 20:55       ` Paul Eggert
2018-08-14 21:03         ` Daniel Colascione
2018-08-14 22:38           ` Paul Eggert
2018-08-15 16:41             ` Eli Zaretskii
2018-08-15 17:16               ` Paul Eggert
2018-08-15 17:47                 ` Eli Zaretskii
2018-08-16  0:29               ` Tom Tromey
2018-08-16 13:26                 ` Eli Zaretskii
2018-08-16 15:43                   ` Daniel Colascione
2018-08-16 16:22                     ` Andreas Schwab
2018-08-19 18:17                     ` Tom Tromey
2018-08-19 19:00                       ` Eli Zaretskii
2018-08-19 19:16                         ` Tom Tromey
2018-08-19 20:23                       ` Stefan Monnier
2018-08-18 10:10                   ` Steinar Bang
2018-08-18 11:31                     ` Eli Zaretskii
2018-08-19 10:00                       ` Robert Pluim
2018-08-19 15:01                         ` Eli Zaretskii
2018-08-19 15:26                   ` Tom Tromey
2018-08-23  0:47                 ` Tom Tromey
2018-08-23 16:48                   ` Eli Zaretskii
2018-08-24 17:54                     ` Tom Tromey
2018-08-24 20:23                       ` Eli Zaretskii
2018-08-24 21:03                         ` Tom Tromey
2018-08-25  6:51                           ` Eli Zaretskii
2018-09-10 11:03                             ` Ergus
2018-09-10 11:15                               ` Robert Pluim
2018-09-10 11:53                                 ` Tom Tromey
2018-09-12 13:37                                   ` Robert Pluim
2018-09-13  4:32                                     ` Tom Tromey
2018-08-16  0:03   ` Tom Tromey
2018-08-16  2:45     ` Eli Zaretskii
2018-08-16 18:07       ` Eli Zaretskii
2018-08-13 13:50 ` T.V Raman
2018-08-13 15:18   ` Tom Tromey
2018-08-13 15:23     ` T.V Raman
2018-08-13 15:15 ` Eli Zaretskii
2018-08-20 21:54   ` John Wiegley
2018-08-13 23:31 ` Richard Stallman
2018-08-13 23:51   ` Tom Tromey
2018-08-16  2:42     ` Richard Stallman
2018-08-15  0:21 ` Clément Pit-Claudel
2018-08-16  0:32   ` Tom Tromey [this message]
2018-08-16  2:14     ` Clément Pit-Claudel
  -- strict thread matches above, loose matches on Subject: below --
2016-12-11 17:37 Nickolas Lloyd
2016-12-12  6:07 ` John Wiegley
2016-12-12 11:51   ` Nickolas Lloyd
2016-12-12 16:45     ` John Wiegley
2016-12-23 17:22       ` Nickolas Lloyd
2016-12-13 22:24   ` Johan Bockgård
2016-12-05 18:16 Burton Samograd
2016-12-05 18:40 ` Eli Zaretskii
2016-12-05 19:32 ` Daniel Colascione
2016-12-05 21:03   ` Burton Samograd
2016-12-06 15:54     ` Lluís

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=877ekri21o.fsf@tromey.com \
    --to=tom@tromey.com \
    --cc=cpitclaudel@gmail.com \
    --cc=emacs-devel@gnu.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. 
Code repositories for project(s) associated with this external index

	https://git.savannah.gnu.org/cgit/emacs.git
	https://git.savannah.gnu.org/cgit/emacs/org-mode.git

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.