Boost logo

Boost-Build :

From: Jürgen Hunold (hunold_at_[hidden])
Date: 2006-04-07 04:15:41


Hi !

Am Freedag 07 April 2006, 08:08 schrieb Vladimir Prus:
> On Friday 07 April 2006 09:58, Stefan Slapeta wrote:
> > during my test runs for intel 9 and cw, bjam comsumes something between
> > 400 and 500 mb of virtual memory (win32 system). I really can't imagine
> > what this memory could be used for, but are there any ways to reduce
> > this hunger?
>
> This would require help from you. Go to the "status" directory and run:
>
> bjam -n -a --v2 intel-win-9.0 cw-9.0 -d+10 > profile
>
> Then, send me the 'profile' file and I'll see if there's something
> particularly wrong there.

I think Stefan is encountering the same memory issues I have reported already
(see Message-Id: <200603221511.26390.hunold_at_[hidden]> )

Stefan, could you try the procedure Volodya proposed this morning ?

He wrote:
> As a quick check, can you go to 'debug.h' and define PROFILE_ENTER and
> PROFILE_EXIT as empty and recompile, and try again. This way we'll be sure
> it's not profiling code itself that cause problems.
>
> After that, I'd go to filent.c, function file_dirscan and wrap some parts
> of it in
>
>  PROFILE_ENTER(FILE_DIRSCAN_PART1)
>  PROFILE_EXIT(FILE_DIRSCAN_PART1)
>
> and maybe this will give us the exact code block that leaks.

I've tried this but got misterious compiler warnings. I think I've lost all my
C-knowledge :-((

Yours,

Jürgen

-- 
* Dipl.-Math. Jürgen Hunold  ! Ingenieurgesellschaft für 
* voice: ++49 511 262926 57  ! Verkehrs- und Eisenbahnwesen mbH  
* fax  : ++49 511 262926 99  ! Lister Straße 15
* hunold_at_[hidden]          ! www.ive-mbh.de

Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk