[Rd] Slow ttests in R-devel

[Prof Brian Ripley]

Are you making use of S4 classes?  If so see the threads on slowness in 
R-patched in the last 48 hours: R-devel will be similar to R-patched was
the last time a branch update was done.

On Wed, 4 Jun 2003, Jeff Gentry wrote:

> I've noticed that some of our Bioconductor code was running drastically
> slower under current R-devel vs. current R-patched - one example is below
> using a ttest.  I have the following snippet of code that demonstrates the
> problem while avoiding "real" code that takes an extremely long time to
> finish on R-devel:
> 
> library(genefilter)
> data(eset)
> eset$cov1
> z <- ttest(eset$cov1, p=0.01)
> gf <- filterfun(z)
> system.time(genefilter(eset, gf))
> 
> (this requires the genefilter & Biobase packages from Bioconductor)
> 
> Using R-1.7.0p I had the following output:

What is that?  We do have 1.7.1 beta now.

> > data(eset)
> > eset$cov1
>  [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2
> > z <- ttest(eset$cov1, p=0.01)
> > gf <- filterfun(z)
> > system.time(genefilter(eset, gf))
> [1] 41.73  2.04 44.45  0.00  0.00
> 
> And using R-devel, I have:
> > data(eset)
> > eset$cov1
>  [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2
> > z <- ttest(eset$cov1, p=0.01)
> > gf <- filterfun(z)
> > system.time(genefilter(eset, gf))
> [1] 80.56  4.20 85.12  0.00  0.00
> 
> Any ideas what might be causing this?
> 
> Thanks
> -Jeff
> 
> ______________________________________________
> R-devel at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
> 

-- 
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