GNU bug report logs - #43269
28.0.50; [feature/native-comp] provide a user feedback on Emacs being native compiled

Previous Next

Full log

Message #38 received at 43269 <at> debbugs.gnu.org (full text, mbox):

From: Andrea Corallo <akrl <at> sdf.org>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: rms <at> gnu.org, arthur.miller <at> live.com, 43269 <at> debbugs.gnu.org
Subject: Re: bug#43269: 28.0.50;
 [feature/native-comp] provide a user feedback on Emacs being native
 compiled
Date: Wed, 09 Sep 2020 18:15:51 +0000

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

>> From: Andrea Corallo <akrl <at> sdf.org>
>> Cc: rms <at> gnu.org,  arthur.miller <at> live.com,  43269 <at> debbugs.gnu.org
>> Date: Wed, 09 Sep 2020 16:32:58 +0000
>> 
>> > Is this with "make" or with "make -jN"? if the latter, what value of N
>> > was used?
>> 
>> $ git clean -xfd && ./autogen.sh && ./configure --without-x
>> --with-nativecomp && time make NATIVE_FAST_BOOT=1 -j16
>> [...]
>> real    4m19.570s
>> user    28m59.958s
>> sys     0m48.797s
>> $
>> 
>> I guess -j1 may even score slightly less user time.
>
> So what do you have there, an i9 CPU?  In any case, 4 min with -j16 is
> quite a lot.

It's Xeon from three yeas ago (8 real cores).  It's 4 mins because our
build is not very parallel.

But that said I think what matters it the total CPU time (here ~30min)
to be compared against the same for the vanilla build (~12min).  This is
about what one would get at -j1.

This indicates a 2.5x.

> And what does NATIVE_FAST_BOOT=1 do? what kind of compiler
> optimizations does it imply?

It's just the current way to say not to compile AoT all the Emacs
distribution but only what goes into the dump.  Is going to to be the
default soon as agreed.

  Andrea

Previous Next

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