On Thursday 03 May 2007 00:53, Han-Wen Nienhuys wrote:
> 2007/5/2, Joe Neeman <[EMAIL PROTECTED]>:
> > >     Well, I have seen that the Lilypond version 2.8.8 can easily be
> > > installed in the local directory, and with that, everything compiles in
> > > a couple of seconds with no fuss.
> >
> > Yes, I can confirm that 2.10 (and 2.11) seem to use a significant amount
> > more memory than 2.8.
>
> AFAICR, 2.8 predates your changes to the pagebreaker; we keep all of
> the scores in memory now, which will signficantly increase memory use.
>  However, it would be nice to know some more useful facts about the
> memory use.
>
> Here are some tips:
>
>  - -ddebug-gc will print statistics of live objects in a file. Look at
> this to see what type of objects are using most memory
>
>  - -dtrace-memory-frequency will sample memory use over time. Look at
> this to see which stages of the program use a lot of memory
>
> Another useful metric is to know when (which version, or better yet:
> which commit) introduced increased memory use.
>
> Unfortunately, I lack the time to investigate this deeply now, and I
> would be grateful if someone did a bit of legwork for me.

Thanks for the tips, I'll look at this after I finish with the page breaker.


_______________________________________________
bug-lilypond mailing list
bug-lilypond@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-lilypond

Reply via email to