[R] speed?

[Prof Brian Ripley]

On Fri, 2 Jun 2006, Roger D. Peng wrote:

> Running on 64-bit per se does not make things faster.  In fact, from my 
> experience it sometimes makes things slower.  The advantage with 64-bit 
> is the extra address space for storing things in memory.

See the R-admin manual for some additional comments.  Certainly until you 
use more than say 300Mb the extra overhead of 64-bit pointers is pure 
overhead. (At some point nearer the address space limit you do GC a lot 
more often on a 32-bit platform.)

> Of course, today's 64-bit chips are all faster than recent 32-bit chips, 
> so you will get a speed improvement, but only because you're getting a 
> better chip.

Not entirely true right now: some of the Pentium Core/Duo Core 32-bit 
chips are competitive.  (Pentium M chips have often been very fast for 
their nomimal clock speeds.)

> -roger
>
> Kerpel, John wrote:
>> The benchmark report is good stuff - I've been wondering about these
>> speed issues recently myself.
>>
>> Has anyone tried something similar on 64-bit Linux or other OS?  I'm
>> contemplating switching to 64-bit Linux if I'll get some dramatic cycle
>> time improvements.  Anyone have any experience with this?
>>
>> Best,
>>
>> John
>>
>> -----Original Message-----
>> From: r-help-bounces at stat.math.ethz.ch
>> [mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of Liaw, Andy
>> Sent: Friday, June 02, 2006 10:01 AM
>> To: ivo welch; r-help at stat.math.ethz.ch
>> Subject: Re: [R] speed?
>>
>> You (and your colleague) might want to have a look at
>> http://www.sciviews.org/benchmark/.  It's a bit dated,
>> but still may be a good starting point.
>>
>> Some months ago some one asked about working on getting
>> R to use the GPU for computation on the R-devel list.
>> Don't know if anything came of it.
>>
>> Cheers,
>> Andy
>>
>> From: ivo welch
>>> dear R wizards:
>>>
>>> while extolling the virtues of R, one of my young
>>> econometrics colleagues told me that he still wants to run ox
>>> because [a] his code is written in it (good reason); [b]
>>> because ox seems to be faster than R in most benchmarks (huh?).
>>>
>>> this got me to wonder.  language speed can't matter much, so
>>> it must be mostly the underlying matrix algebra by now.  I
>>> presume that nowadays most of the plain matrix operation
>>> speed depends primarily on which hardware features the
>>> library accesses.  (The basic algorithms aren't so different,
>>> so even though the algorithm may have mattered a long time
>>> ago, they are probably pretty similar nowadays. hmmm...maybe
>>> matrix inversion still is different, but multiplication and
>>> adding should not be.)
>>>
>>> On x86 architecture, I believe there is a hierarchy in terms
>>> of raw processing power:  FPU < SSE* < GPU.
>>>
>>> is it even possible to use the GPU now for math processing
>>> (linux or windows), and specifically in R?
>>>
>>> assuming I compile everything with the proper SSE flags and atlas, is
>>> SSE* fully taken advantage of?
>>>
>>> regards,
>>>
>>> /ivo
>>>
>>> ______________________________________________
>>> R-help at stat.math.ethz.ch mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-help
>>> PLEASE do read the posting guide!
>>> http://www.R-project.org/posting-guide.html
>>>
>>>
>>
>> ______________________________________________
>> R-help at stat.math.ethz.ch mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide!
>> http://www.R-project.org/posting-guide.html
>>
>> ______________________________________________
>> R-help at stat.math.ethz.ch mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
>>
>
>

-- 
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