[R] re: smp in Linux

[A.J. Rossini]

>>>>> "TL" == Thomas Lumley <thomas at biostat.washington.edu> writes:

    TL> On Wed, 10 Mar 1999, Kenneth Nordstrom wrote:
    >> A question to all you R-gurus:
    >>
    >> Can R (or S-plus, for that matter) make efficient use of
    >> multiple Intel Processors running under Linux (within the same
    >> PC, not over a net)?  With the release of the new 2.2 kernel,
    >> this would seem a interesting and cost-efficient way of
    >> boosting the computational power of Intel/Linux platforms when
    >> using R (or S-plus).

S4 has a threaded version (which is what Splus 5.0 is based on).  It's
a bit broken, and definitely alpha-quality.  When it works, it's
rather nice in an SMP environment.

    TL> There are possibilities for parallelisation short of making R
    TL> fully threaded.  One more-or-less straightforward one would be
    TL> to replace the matrix operations with efficient parallel
    TL> code. This wouldn't help a lot since R doesn't spend that much
    TL> time doing matrix operations.  Another possibility would be to

One thing I'm scheduled to try out is compiling R using the portland
group compilers (www.pgroup.com, I think?), which optimize for SMP and
parallel LAN environments.  These aren't free, but some people have
seen decent improvements with them.

I'm not sure that it'll help very much, though, unless a good deal of
changes are made in the configuration to use the right libraries (it's
also not high on my To-Do list right now -- ask again in a few
months).

best,
-tony

-- 
A.J. Rossini
UW Biostatistics & Center for AIDS Research 
206-543-1044 / 206-720-4282                 
rossini at biostat.washington.edu              
http://www.biostat.washington.edu/~rossini/ 
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._