[darcs-devel] Repository.writePatch (issue80)

[Aggelos Economopoulos]

On Sunday 15 January 2006 21:24, Jason Dagit wrote:
> On Jan 15, 2006, at 7:34 AM, Aggelos Economopoulos wrote:
> > On Sunday 15 January 2006 05:04, Jason Dagit wrote:
> >> Peak RES (as measured by top)
> >> orig: 940MB
> >> no-reread: 950MB
> >>
> >> Peak VIRT (as measured by top)
> >> orig: 1756MB
> >> no-reread: 1289MB
> >>
> >> When I realized that orig had consumed more time and space than no-
> >> reread I killed it so I'm not sure where it would have peaked or how
> >> much time it would have taken to reach that peak.
> >
> > Err, according to your numbers, Peak RSS was larger for no-reread,
> > so orig did
> > _not_ consume greater max memory up to that point (I'm not claiming
> > that it
> > wouldn't). Peak VIRT is generally not indicative of actual memory
> > usage.
>
> Well, if you want to ignore VIRT, then the two versions come with in
> 10megs of each other out of almost 1GB, and yet orig was still
> running after 5 hours.  I don't see how that can be preferred.

Never said I prefer one behaviour over the other. All I'm saying is that since 
you obviously care about improving darcs in this case, *you* want to ignore 
VIRT and make proper measurements.

> I 
> would also imagine the reason both versions don't use more RSS is
> because I only have 1GB of ram.

How is peak RSS meaningful then? You might want to try using the extra fields 
in top and maybe take a look in /proc/$pid (assuming you're running linux) if 
you want a more accurate view of things (wrt. to memory usage). Or try 
Juliusz's suggestion.

HTH,
Aggelos