[Rd] Single vs. dual CPUs

[Peter Dalgaard]

"Milton Lopez" <mlopez at iattc.org> writes:

> I've posted this earlier and have not heard much so far. I'd really appreciate any guidance on this as we are about to order new hardware.
> 
> We are buying Dell workstations with Red Hat Linux and 64-bit Xeon
> CPUs to run R. We could add a second processor to each system, or
> buy slightly faster single CPU systems. Is it possible to make a
> generalized statement as to what kind of performance improvement we
> would see with a single vs. dual processors when running R on these
> systems?

Well, if you ask that way, the answer is probably no...

It depends on the usage pattern. If you run multiple CPU-bound
processes in parallel without too much coordination (parallel make is
a good example, simulations another), then you get close to double up
from a dual. For a single R process, you can get something like 40%
improvement in large linear algebra problems, using a threaded ATLAS.
For other problems the speedup is basically nil. There is some
potential in threading R or (much easier) some of its vector
operations, but that is not even on the drawing board at this stage.

Also, these days you might want to consider another factor: noise.
Duals tend to be server machines with little emphasis on quietness,
where the single-CPU machines have heatpipes and whatnot. 

-- 
   O__  ---- Peter Dalgaard             Øster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark          Ph:  (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)                  FAX: (+45) 35327907