GNU bug report logs - #41194
28.0.50; [feature/native-comp] fibn benchmark exhausts memory

Previous Next

Full log

Message #20 received at 41194 <at> debbugs.gnu.org (full text, mbox):

From: Kévin Le Gouguec <kevin.legouguec <at> gmail.com>
To: Andrea Corallo <akrl <at> sdf.org>
Cc: 41194 <at> debbugs.gnu.org
Subject: Re: bug#41194: 28.0.50; [feature/native-comp] fibn benchmark
 exhausts memory
Date: Sun, 14 Jun 2020 17:19:29 +0200

[Message part 1 (text/plain, inline)]

Andrea Corallo <akrl <at> sdf.org> writes:

> 5a55a845a7 * Implement 'maybe_gc_or_quit' to allow correct GC in compiled Lisp
> 34ed9d2498 * Introduce latches
>
> are fixing here this issue, now running fibn at speed 2 the memory
> consuptions stays constant.

Alright, feedback time!

Commits:
- master: 82a632edc8b80bf16d9b9f205474bf9724b084c0
- feature/native-comp: 5a55a845a7c426e82e8a6a6d02bc4a39992871e3


The Good News
=============

- I can compile the branch without adding files to the default
  blacklist,

- compilation (-j1) now only takes less than 6 hours instead of 3 days
  (compared with 1.5 hour on master),

- at no point during compilation did the system swap; in fact, I could
  probably have risked a -j2 build,

- elisp-benchmarks doesn't show any obvious performance regression
  (… though see Bad News).

For details on memory usage during compilation, see attached memory
profile (to compare with bug#41077#28).

[compilation-memory.pdf (application/pdf, attachment)]

[Message part 3 (text/plain, inline)]

Few files take more than 3 minutes to compile now:

ELC+ELN emacs-lisp/comp.el      0:09:49
ELC+ELN org/org.el              0:03:39
ELC+ELN ../lisp/window.el       0:03:16
ELC+ELN emacs-lisp/bytecomp.el  0:03:16

For details on performance, see attached tables for master and
feature/native-comp, respectively (to compare with bug#41077#40).

[perfs.org (text/x-org, attachment)]

[Message part 5 (text/plain, inline)]

The Bad News
============

- fibn still exhausts memory,
- now pidigits does too.

[fibn.pdf (application/pdf, attachment)]

[pidigits.pdf (application/pdf, attachment)]

[Message part 8 (text/plain, inline)]

I've only tried the default comp-speed (2) so far.  I've attached the
results of M-x disassemble on the functions in fibn.el and pidigits.el.

[elb-eliminate-digit.s (text/x-asm, attachment)]

[elb-extract-digit.s (text/x-asm, attachment)]

[elb-fibn.s (text/x-asm, attachment)]

[elb-fibn-entry.s (text/x-asm, attachment)]

[elb-next-term.s (text/x-asm, attachment)]

[elb-pidigits.s (text/x-asm, attachment)]

[elb-pidigits-entry.s (text/x-asm, attachment)]

[Message part 16 (text/plain, inline)]

Let me know what I can do to help diagnose this further.  Again, thank
you for the time spent on this issue!

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.