[R] fastest R platform
M. Edward (Ed) Borasky
znmeb at aracnet.com
Mon Apr 9 20:18:54 CEST 2001
On Mon, 9 Apr 2001, Prof Brian Ripley wrote:
> It has never been clear to me in this thread what OS the PC concerned is
> running. Not all OSes can run the profiler, which is very useful but
> coded in rather specific way. Implementing profiling on Windows is
> an open project.
I thought he said "Windows ME" :-(. If so, maybe this isn't an R issue but a
Microsoft one. I had Windows ME for six months before giving up on it. The
choice was between going back to Windows 98 SE, going forwards to Linux or to
Windows 2000 Professional. I ended up with a dual-boot Red Hat 7.0 and Win2K.
The bulk of my work is done with Win2K because my financial database only does
Windows. The music stuff I do mostly in Linux now, though. Speaking of which,
is anyone interested in using R for "sonification"? I'm toying with the idea
of integrating R with the public-domain "sfront" package, which is in pure C
and runs on both Linux and Windows. This would give R the ability to add
"sonic displays" to its graphic capabilities.
> > On platforms: R performs at similar speed between Windows and Linux, and
> > my limited comparisons between Intel and SPARC machines suggest that the
> > SPECint rating will give a reasonable ballpark estimate of speed across
> > platforms. This means you probably don't want to shift off Intel for
> > speed reasons unless you have a LOT of money.
> And before you do you probably want to examine the possibilities of
> parallelization on cheap Intel i686 boxes. I know R is not parallel,
> but most statistical tasks are parallelizable.
Most *large* tasks are parallelizable; most small ones are not *effectively*
parallelizable. I suspect what the original poster has is a medium task --
something with enough structure to be tunable for a single-processor but
paralellizable only in the sense that, if he's doing multiple independent runs
varying some parameter, he can do those independent runs on separate machines.
If he truly is running Windows ME, I'd recommend spending another
$ 500 or $ 1000 on a new system, putting Linux on it and tuning his job
there with R profiling.
Then he can run the tuned code on the Windows system.
znmeb at aracnet.com (M. Edward Borasky) http://www.aracnet.com/~znmeb
If there's nothing to astrology, how come so many famous men were born
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
More information about the R-help