[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Optimising Octave compilation time
From: |
Richard Campbell |
Subject: |
Re: Optimising Octave compilation time |
Date: |
Fri, 18 Feb 2011 14:47:09 -0500 |
On Feb 18, 2011, at 2:28 PM, Jordi GutiƩrrez Hermoso wrote:
> So you really compile all of the Octave sources in 2 minutes,
> Campbell? Wow. Maybe that's why I'm the only one who complains about
> having to regularly recompile Octave when someone decides that we need
> to have two spaces instead of one after every full stop. I've kinda
> solved this problem with ccache, which does tend to cut down my
> compilation times.
>
> 2011/2/18 Richard Campbell <address@hidden>:
>> I've never encountered race conditions with make - that would be a
>> problem with make, though, not the Octave sources, right?
>
> I don't know. I've never tried with Octave, but I have seen it
> elsewhere. I don't often spawn multiple threads with Probably an
> improperly autotooled project. Then again, I don't often try to spawn
> multiple compilation threads on this laptop. I'm afraid of overheating
> it.
>
> - Jordi G. H.
https://mailman.cae.wisc.edu/pipermail/octave-maintainers/2011-February/022701.html
This was with octave-3.3.5x. A few days later when I checked out the Mercurial
sources the first time, I remember them building in 2 minutes 1 second with
gcc. I didn't try clang.
I'm trying to reproduce it now, but ./autogen.sh is hanging at "Cloning into
gnulib..." and I see no network activity. What gives?
It helps to have 24 logical cores, but even if I only had one, it seems that it
would take < 45 minutes, at least with the Mercurial sources. And there aren't
any Macs on the market capable of running Snow Leopard that have less than 2
cores, so I'm thinking more like 25 minutes, worst case?
Campbell