[Bf-committers] Meeting minutes, nov 7, 2010

Campbell Barton ideasman42 at gmail.com
Mon Nov 8 12:46:40 CET 2010


On Mon, Nov 8, 2010 at 7:51 AM, Toni Alatalo <antont at kyperjokki.fi> wrote:
> On ma, 2010-11-08 at 07:20 +0000, Campbell Barton wrote:
>> On Sun, Nov 7, 2010 at 5:13 PM, Ton Roosendaal <ton at blender.org> wrote:
>> > - Campbell proposes to ditch old Makefiles, Ton protests, but will
>> > test Cmake in due time. Andrea mentions Cmake to MSVC has a nasty bug,
>> For the record (ie. in-case Ton changes his mind). just did some tests
>> with CMake Makefiles vs NAN-Makefiles.
>> Single threaded build:
>>  CMake: 4min, 18 seconds.
>>  NAN-Makefiles: 7min, 45seconds.
>
> AFAIK what Ton has traditionally been concerned with is little
> incremental builds after one little change. Which is of course common
> when developing. Used to take some 5-15secs on a G3 mac laptop IIRC :)
> And much longer with SCons unless doing the quickie-thing.
>
> So another thing to test would be doing one change in one file and
> seeing how quickly cmake makefiles give a new blender after that ..
> based on my experience in another cmake using project should be quite
> quick.
>
> ~Toni

I find cmake is fast enough not to need make quicky, even though I
setup a wrapper script to give make-quicky functionality, though on a
slower system, make quicky could still be useful.

Comparison between cmake and make when there is nothing to rebuild:
nan-makefiles: 10.837s
cmake's makefiles: 1.05sec

Rebuild with 1 file change (drawobject.c)
  nan-makefiles: 13.028s
  nan-makefiles quicky: 3.134s
---
  cmake's makefiles: 3.147s
  cmake's makefiles quicky: 2.715s

This is with most options enabled, except buildinfo, collada and
cmake's WITH_INSTALL

I normally disable jack, ffmpeg, sdl, openal, bge etc. a rebuild with
a single change for is under 1.5 seconds.

cpu id - AMD Phenom(tm) II X6 1055T Processor
-- 
- Campbell


More information about the Bf-committers mailing list