> On Monday, 22 May 2000, Han-Wen Nienhuys writes:
> >
> > I added some more tweaks to .52 in order to suppress memory
> > usage. Could the large-score people check if .53 fares better than .52?
>
> Some crude statistics, memory usage and user time of the Coriolan:
>
> 1.3.7 72.844M
> 1.3.9 78.039M
> 1.3.28.hwn1 99.086M
> 1.3.50 117.102M 1904.58s
> 1.3.51 116.273M 998.76s
> 1.3.52 116.758M 792.12s
> 1.3.53 118.605M 876.36s
I get fairly similar results for the Gade score, i.e.
the processing time was halfed from .50 to .51 and
the memory consumption has stayed about the same in the
latest releases. Actually, in my case the processing
time decreased slightly from .52 to .53.
Jan, how did you estimate the memory usage?
/Mats