GNU bug report logs - #45705
[feature/native-comp] Excessive memory consumption on windows 10

Previous Next

Full log

Message #26 received at 45705 <at> debbugs.gnu.org (full text, mbox):

From: Andrea Corallo <akrl <at> sdf.org>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: edouard.debry <at> gmail.com, 45705 <at> debbugs.gnu.org
Subject: Re: bug#45705: [feature/native-comp] Excessive memory consumption
 on windows 10
Date: Fri, 08 Jan 2021 22:02:38 +0000

Eli Zaretskii <eliz <at> gnu.org> writes:

>> From: Andrea Corallo <akrl <at> sdf.org>
>> Cc: edouard.debry <at> gmail.com, 45705 <at> debbugs.gnu.org
>> Date: Fri, 08 Jan 2021 15:50:28 +0000
>>
>> consumptions of few GiB is something I've seen more then once for long
>> standing sessions.  You might be right in this being a memory leak,
>> indeed I've no prove of that (I think we have none for the other
>> direction either).
>
> I'm not yet worried about memory leaks, I'm more worried that there's
> no memory leaks, and instead using libgccjit indeed requires such
> large memory amounts.  Are you sure this is not the case?

I see,

if we are interested in comparing the memory footprint of using shared
libraries for functions vs stock byte-code I think we can "statically"
compare two sessions after the startup.

I've compiled current native-comp with and without --with-nativecomp
repeating the experiment with and without X.  These are the data-points:

  |         | --without-x | --without-x --with-nativecomp |      |
  |---------+-------------+-------------------------------+------|
  | -Q      | 49M         | 92M                           | 1.9x |
  | my-conf | 92M         | 179M                          | 1.9x |
  
  
  |         |      | --with-nativecomp |      |
  |---------+------+-------------------+------|
  | -Q      | 536M | 756M              | 1.4x |
  | my-conf | 585M | 1453M             | 2.4x |

So yes shared are using considerably more memory, I think this is
expected as also the file footprint suggests native code is less dense
that byte-code (actually with a quite similar relative results).

Indeed *with use the delta should decay as data are the same and there's
no difference in its representation*, so this picture should be more on
the worst case side than on the average.

Also if we want to see a positive side, multiple Emacs sessions will
share the majority the pages allocated for the shared libraries.

  Andrea

Previous Next

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