[R] R - How to "physically" Increase Speed

Henrik Bengtsson hb at biostat.ucsf.edu
Tue Oct 22 07:36:51 CEST 2013

On Mon, Oct 21, 2013 at 4:12 PM, Alexandre Khelifa
<akhelifa at logitech.com> wrote:
> Hi,
> My name is Alexandre Khelifa and I have been using R at my work for about 2
> years. I have been working on a project when we do Monte Carlo Simulation
> and it involves a lot of calculations.
> I am currently using R x64.3.0.1 and used to work on a 4GB machine.
> However, the calculation time was very long (about 2 weeks), and the IT
> team and I decided to add more memory and to make it a 8GB virtual machine.
>  I also added the following line in my code:
> *
> *
>    - *options(java.parameters = "-Xmx8192m") *
> I re-did the same calculations (4GB vs. 8GB) but did not see a significant
> increase in the calculation time, so I was wondering if I did anything
> wrong and/or what would be the best solution to increase this time.

If you and your team travel Earth-to-Space round trip at 87% the speed
of light, your computations would *physically* become roughly twice as
fast (you must leave computer behind).  Alternatively, look at
help(package="parallel").  But what really makes a difference is to
find and replace bottle necks in your code by profiling it, cf.
help("Rprof").  I'd go with the latter if you haven't already done so.


> Thanks a lot for your help and more generally for building such an amazing
> (and free) tool.
> Let me know if you have any questions.
> Regards,
> Alexandre Khelifa
>         [[alternative HTML version deleted]]
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

More information about the R-help mailing list