[Scons-dev] SCons speedup and profiling results...

[rupert THURNER]

On Thu, Sep 26, 2013 at 8:44 AM, Dirk Bächle <tshortik at gmx.de> wrote:

> Hi Gary,

>

> On 26.09.2013 02:08, Gary Oberbrunner wrote:

>>

>>

>>

>> [...]

>>

>>

>> I think this is excellent work!  Solid analysis.  I know there's been some

>> thought given to caching subst() before; it's trickier than one might think

>> but in many cases it should work, and it definitely speeds things up.  I'm

>> also impressed by a 30% memory reduction -- interested to hear how that

>> comes out.

>>

>

> it will definitely take some more time. Not only subst() is quite tricky,

> but getting cyclic dependencies out of the way for garbage-collecting Nodes

> means a lot of fiddling. But there should be some light at the end of the

> tunnel soon....

>

> By the way: Does anybody know of a use case for having more than one FS

> during a build or interactive mode? I've wondered many times now, why on

> earth we keep a reference to the current DefaultFS in each Node? After all I

> don't see any place in the source code where I could initialize a second

> one, or one that's different from the Default FS...

>

>

>> Have you seen the speed and memory tests the buildbots used to run?  Not

>> sure what their status is these days but perhaps your tests could be

>> integrated into them at some point.

>

>

> Yes, I considered using them. But as far as I know, they rely on everything

> that's being run to be a packaged archive and you always have to write some

> sort of config file. That's what put me off a little bit, and so I stuck to

> my own scripts/packages and developed them further...


what are the main things which make waf, and wonderbuild so much
faster than scons? a test is here, but already a little outdated:
http://retropaganda.info/~bohan/work/psycle/branches/bohan/wonderbuild/benchmarks/time.xml


rupert.