unofficial mirror of bug-gnu-emacs@gnu.org 
 help / color / mirror / code / Atom feed
From: "Kévin Le Gouguec" <kevin.legouguec@gmail.com>
To: Andrea Corallo <akrl@sdf.org>
Cc: 41077-done@debbugs.gnu.org
Subject: bug#41077: [feature/native-comp] virtual memory exhausted
Date: Mon, 11 May 2020 11:12:47 +0200	[thread overview]
Message-ID: <87d07avnqo.fsf@gmail.com> (raw)
In-Reply-To: <xjfimh34ep7.fsf@sdf.org> (Andrea Corallo's message of "Sun, 10 May 2020 22:17:56 +0000")

Ah, right; I don't know what I was thinking.  I guess I assumed that the
benchmarks covered Emacs subroutines, so not compiling the benchmark
sources would not have a big impact.  I got that completely backward 🤦

Andrea Corallo <akrl@sdf.org> writes:

> Other option maybe quicker: you can manually native compile the
> benchmarcks and the native emacs instance should load them.
>
> You will see that when a .eln is loaded the message is a little
> different to indicate that.

I went with this option, using M-: (native-compile "foo.el").  I had one
last surprise: the native-compiled fibn test exhausts the laptop's
memory (2GB RAM + 2GB swap).  I recompiled fibn.eln setting comp-speed
to 1 and 0, but the result was the same.

Here are the results: except for pidigits, the improvements are quite
noticeable :)  I'll post a write-up (giving my hardware specs and
computing the speedups) to emacs-devel when I find the time.

* Results

** master

  | test           | non-gc avg (s) | gc avg (s) | gcs avg | tot avg (s) | tot avg err (s) |
  |----------------+----------------+------------+---------+-------------+-----------------|
  | bubble-no-cons |         122.06 |       0.22 |       4 |      122.28 |            0.43 |
  | bubble         |          49.22 |      26.59 |     491 |       75.81 |            0.18 |
  | fibn-rec       |          62.00 |       0.00 |       0 |       62.00 |            0.00 |
  | fibn-tc        |         133.97 |      87.95 |    1520 |      221.91 |            0.36 |
  | inclist        |         174.41 |       0.06 |       1 |      174.47 |            0.00 |
  | listlen-tc     |          68.04 |       0.00 |       0 |       68.04 |            0.02 |
  | pidigits       |          56.97 |      36.60 |     605 |       93.57 |            0.53 |
  |----------------+----------------+------------+---------+-------------+-----------------|
  | total          |         666.67 |     151.42 |    2621 |      818.08 |            0.79 |

** feature/native-comp

  | test           | non-gc avg (s) | gc avg (s) | gcs avg | tot avg (s) | tot avg err (s) |
  |----------------+----------------+------------+---------+-------------+-----------------|
  | bubble-no-cons |          24.53 |       0.22 |       4 |       24.75 |            0.02 |
  | bubble         |          17.83 |       5.79 |     100 |       23.62 |            0.27 |
  | fibn-rec       |          18.63 |       0.00 |       0 |       18.63 |            0.01 |
  | fibn-tc        |          67.39 |      86.45 |    1519 |      153.84 |            0.07 |
  | inclist        |          10.99 |       0.06 |       1 |       11.05 |            0.04 |
  | listlen-tc     |           0.82 |       0.00 |       0 |        0.82 |            0.00 |
  | pidigits       |          54.03 |      36.84 |     605 |       90.87 |            0.69 |
  |----------------+----------------+------------+---------+-------------+-----------------|
  | total          |         194.22 |     129.35 |    2229 |      323.57 |            0.75 |

Again, thanks for your help!





  reply	other threads:[~2020-05-11  9:12 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-05-04 15:08 bug#41077: [feature/native-comp] Segfaults when compiling ELC+ELN Kévin Le Gouguec
2020-05-04 16:31 ` Andrea Corallo
2020-05-04 20:57   ` Andrea Corallo
2020-05-04 21:05   ` Kévin Le Gouguec
2020-05-04 21:15     ` Andrea Corallo
2020-05-06 14:15       ` bug#41077: [feature/native-comp] virtual memory exhausted (was: bug#41077: [feature/native-comp] Segfaults when compiling ELC+ELN) Kévin Le Gouguec
     [not found]         ` <xjfftcd5chc.fsf@sdf.org>
2020-05-06 20:12           ` bug#41077: [feature/native-comp] virtual memory exhausted Andrea Corallo
2020-05-10 14:26             ` Kévin Le Gouguec
2020-05-10 15:02               ` Andrea Corallo
2020-05-10 22:04                 ` Kévin Le Gouguec
2020-05-10 22:17                   ` Andrea Corallo
2020-05-11  9:12                     ` Kévin Le Gouguec [this message]
2020-05-11 10:00                       ` Andrea Corallo

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

  List information: https://www.gnu.org/software/emacs/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87d07avnqo.fsf@gmail.com \
    --to=kevin.legouguec@gmail.com \
    --cc=41077-done@debbugs.gnu.org \
    --cc=akrl@sdf.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 public inbox

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

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).