[R-SIG-Mac] Mac as R platform

[Jan de Leeuw]

There are various reasons why the Mac is not the best option if you
want raw speed.

-- The current dual core Intel chips used in Macs are mainly developed  
for portables (i.e. they emphasize power saving). The next generation,
available in July/August, will do a better job and will be used in the
new towers.

-- The G5 is not designed to do fast floating point, and generally does
not do a very good job at it.

-- OS X is handicapped by the Mach microkernel parts. This is becoming
more and more obvious in webserving, database management, and  
distributed
computing. Rumor has it that Apple is working on both replacing the
kernel and the HFS+ file system.

For my work this is all pretty irrelevant, but many others care. And  
with
good reason.

I do think, however, that the Mac, expecially the MacBook Pro,  is a  
pretty good
development machine -- especially my MacBook Pro, since it now  
painlessly virtualizes
XP Pro, Debian, and Solaris 10 as well.


On May 8, 2006, at 23:11 , Simon Urbanek wrote:

> Jason,
>
> On May 5, 2006, at 12:07 PM, Jason Liao wrote:
>
>> Good morning. I am considering buying a new computer to develop an
>> R package for multi-dimensional Monte Carlo simulation. One
>> requirement is to call, from R, a Fortran 90 program about 1000
>> lines of code. A  C++ version of that program is also available if
>> desired.  Will the new Core Duo Mac be a good development platform
>> in the sense that I will not need to struggle with too many
>> technicalities in linking R with the compiled code? I am asking
>> this because I saw that the binary Mac R distribution in fact has
>> compilers bundled in it. I currently work mostly on Windows and was
>> told that I will have a lot to download and configure if developing
>> on Windows.
>>
>
> in general I think Macs are pretty painless in terms of developing
> packages for R. Most tools come right from Apple with the operating
> system. To complement that the fortran compiler supplied with R (GNU
> Fortran 95) supports F90, so you should be safe (although I'm not
> sure how versatile such package would be). At the development stage
> you shouldn't worry about universal binaries, either. Admittedly, I
> develop even Windows programs on my Mac, because I find it easier to
> cross-compile for Windows on a Mac than to do it in the horrible
> Windows GUI, but that's another story ;). I don't know if that helps,
> but those are just my 2 cents.
> However, if you're interested in raw speed (i.e. you actually want to
> run the simulations), you may be better off using dual-core AMD
> Opterons with Linux or Quad-G5 PowerMacs - the Intel CPUs are not
> necessarily fast (although Apple tries to make you believe
> otherwise ;)).
>
> Cheers,
> Simon
>
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>

===
Jan de Leeuw; Distinguished Professor and Chair, UCLA Department of  
Statistics;
Editor: Journal of Multivariate Analysis, Journal of Statistical  
Software
US mail: 8125 Math Sciences Bldg, Box 951554, Los Angeles, CA 90095-1554
phone (310)-825-9550;  fax (310)-206-5658;  email: deleeuw at stat.ucla.edu
.mac: jdeleeuw ++++++  aim: deleeuwjan ++++++ skype: j_deleeuw
homepages: http://gifi.stat.ucla.edu ++++++ http://www.cuddyvalley.org
   
------------------------------------------------------------------------ 
-------------------------
           No matter where you go, there you are. --- Buckaroo Banzai
                    http://gifi.stat.ucla.edu/sounds/nomatter.au