[R] Odd timing behaviour in reading a file

[Prof Brian Ripley]

On Sun, 7 Aug 2005 Glenn.Stone at csiro.au wrote:

> Thanks, very helpful.
>
> Is there some way to adjust those GC triggers in advance?

Almost: that is what the statrtup flags such as --min-vsize help do.

> -----Original Message-----
> From:	Prof Brian Ripley [mailto:ripley at stats.ox.ac.uk]
> Sent:	Thu 8/4/2005 4:15 PM
> To:	Stone, Glenn (CMIS, North Ryde)
> Cc:	r-help at stat.math.ethz.ch
> Subject:	Re: [R] Odd timing behaviour in reading a file
> Please see the gcFirst argument to system.time, which you should set to
> TRUE for such timings.  Your second run is paying to GC the results of the
> first, most likely.
>
> Beyond that, R adjusts its GC triggers based on usage, and when you first
> start using large objects the trigger levels will grow and generally
> things will speed up.  Set gcinfo(TRUE) to watch what is happening.

[...]

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595