Re: [Boost-testing] Very strange regression results for borland

From: AlisdairM (alisdair.meredith_at_[hidden])
Date: 2007-02-06 03:10:07

Roland Schwarz wrote:

> Bronek Kozicki wrote:
> > I noticed this before - it's bjam consuming huge amounts of memory, 
> > proportional to number of selected toolsets
> 
> This would also be an explanation, why I observed an
> over-proportional increase of total run time.
> 
> Adding two more compilers (5->7) more than doubled the time
> (12h->25h).
> 
> Roland

not sure if it helps, but I also have a patch on mainline to reduce a
large number of Borland warnings - running on just 2 Borland compilers
my log file reduced from 133Mb to 13Mb!  I have not applied to rc
though, unless I hear from Thomas.  (it is low risk, protected by
Borland #if guards, and should not affect code generation)

-- 
AlisdairM