R help - Mar 2009 - recommended computing server for R (March 2009)?

dear r-experts:

I need to speed up my monte-carlo simulations. my code is written in R (and  
it was also the cause of my many questions here over the last few days). my  
code is almost all matrix/vector algebra on panel data  
sets---long-difference, fixed-effects, blundell-bond, etc.. the data set is  
about 10MB, so 1GB per CPU core should be plenty for my operations, and  
with $10/GB of DRAM, this is no longer a bottleneck. For my application,  
parallelism is a given, since most of it is monte-carlo simulations. (I  
guess the diametrically opposite need would be when one cannot parallelize,  
in which case the recommendations would be quite different.)

My operating system will probably be ubuntu. (I also run a little of it on  
an OSX Mac Pro I own.)

I want to use an Intel/AMD system with a prebuilt R executable. I do not  
want to fiddle (too much) with building R myself, unless it is real easy  
and makes a real speed difference. I wish I could ask R to load something  
exotic like CUDA, but I presume that this is not yet ready for prime-time.  
PS3 is probably silly, too. in fact, if I am not mistaken (and I may well  
be), R pre-built does not even take advantage of SSE3 out-of-the-box.

software-wise, is there anything unusual that I should heed, or should I  
just pick of R 2.8.1 from the CRAN archives and be done with it?

now, I also have to make some simple hardware decisions. Right now, a  
dual-socket quad-core AMD opteron shanghai 2.3GHz system seems cheap.  
$174/CPU + $70/motherboard. is there a system that dominates this in terms  
of $/MFlops? (I presume the fact that core i7 has threads is irrelevant to  
R.) I am not trying to ignite a flame-war---in fact, I don't care about any
other features that AMD or Intel or anything might have for this particular  
computer. Other needs may warrant different choices.

Any other thoughts would be appreciated. Although you can just email them  
to me, I presume that this question has enough interest to others that  
posting it is ok.

regards,

/ivo welch

	[[alternative HTML version deleted]]

On 28 March 2009 at 13:37, ivowel at gmail.com wrote:
| I need to speed up my monte-carlo simulations. my code is written in R (and  
| it was also the cause of my many questions here over the last few days). my  
[...]
| with $10/GB of DRAM, this is no longer a bottleneck. For my application,  
| parallelism is a given, since most of it is monte-carlo simulations. (I  
[...]
| My operating system will probably be ubuntu. (I also run a little of it on  
| an OSX Mac Pro I own.)

One thing you could consider is renting the compute hours from the cloud:

    http://aws.amazon.com/ec2

as EC2 now has a choice of Debian and Ubuntu (among others) and Debian /
Ubuntu have R and Open MPI work out of the box.  Examples as in my 'Intro to
High Performance Computing with R' tutorials from UseR and the BoC [ google
for the pdf slides if interested ] should apply 'as is', you don't
need to
fiddle with (physical) hardware and can scale up CPU resources as needed.

Dirk

-- 
Three out of two people have difficulties with fractions.

A SAS spokesperson has confirmed to this blog that they have invested
in the R –Core project to help build next generation algorithms .
Details are sketchy but indications of some shift on cloud hosted SAS
,called SaaS are emerging.Also includes some details on Jim Davis ,SVP
SAS marketing''s statement on BI and Anne Milley having a new
assignment within SAS Institute.

Read more here -
http://www.decisionstats.com/2009/04/sas-institute-invests-in-r-project/



ivowel wrote:
> 
> dear r-experts:
> 
> I need to speed up my monte-carlo simulations. my code is written in R
> (and  
> it was also the cause of my many questions here over the last few days).
> my  
> code is almost all matrix/vector algebra on panel data  
> sets---long-difference, fixed-effects, blundell-bond, etc.. the data set
> is  
> about 10MB, so 1GB per CPU core should be plenty for my operations, and  
> with $10/GB of DRAM, this is no longer a bottleneck. For my application,  
> parallelism is a given, since most of it is monte-carlo simulations. (I  
> guess the diametrically opposite need would be when one cannot
> parallelize,  
> in which case the recommendations would be quite different.)
> 
> My operating system will probably be ubuntu. (I also run a little of it on
> an OSX Mac Pro I own.)
> 
> I want to use an Intel/AMD system with a prebuilt R executable. I do not  
> want to fiddle (too much) with building R myself, unless it is real easy  
> and makes a real speed difference. I wish I could ask R to load something  
> exotic like CUDA, but I presume that this is not yet ready for prime-time.
> PS3 is probably silly, too. in fact, if I am not mistaken (and I may well  
> be), R pre-built does not even take advantage of SSE3 out-of-the-box.
> 
> software-wise, is there anything unusual that I should heed, or should I  
> just pick of R 2.8.1 from the CRAN archives and be done with it?
> 
> now, I also have to make some simple hardware decisions. Right now, a  
> dual-socket quad-core AMD opteron shanghai 2.3GHz system seems cheap.  
> $174/CPU + $70/motherboard. is there a system that dominates this in terms
> of $/MFlops? (I presume the fact that core i7 has threads is irrelevant to
> R.) I am not trying to ignite a flame-war---in fact, I don''t care
about
> any  
> other features that AMD or Intel or anything might have for this
> particular  
> computer. Other needs may warrant different choices.
> 
> Any other thoughts would be appreciated. Although you can just email them  
> to me, I presume that this question has enough interest to others that  
> posting it is ok.
> 
> regards,
> 
> /ivo welch
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help@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.
> 
> 
-- 
View this message in context:
http://www.nabble.com/recommended-computing-server-for-R-%28March-2009%29--tp22756946p22824049.html
Sent from the R help mailing list archive at Nabble.com.

	[[alternative HTML version deleted]]