[R-sig-hpc] doSNOW + foreach = embarrassingly frustrating computation

A.J. Rossini blindglobe at gmail.com
Tue Dec 21 22:38:41 CET 2010

SNOW shouldn't be complex -- not if people do the right thing and
appropriately add a backend (i.e. similar to what Revolution did for
NWS).  If you like to roll your own and target your own architecture,
fine.  And after someone else finds it inefficient, history will
repeat itself rather than learning.

SNOW provides portability, which MPI doesn't provide.  Portability and
reproducibility are important, even if people think that simplicity
and ease of use are wonderful things.

I'll point out the usual analogy,
http://www.jwz.org/doc/worse-is-better.html  for why SNOW won't win
and worse is better.

On Tue, Dec 21, 2010 at 8:18 PM, Brian G. Peterson <brian at braverock.com> wrote:
> On 12/21/2010 11:38 AM, Marius Hofert wrote:
>> Any ideas?
> I wonder  why you are adding the complexity of SNOW on top of MPI instead of
> using doMPI.  There are several good examples and vignette on calling doMPI
> via mpirun, and a discussion of what side effects you can expect and plan
> for.
> Regards,
>  - Brian
> --
> Brian G. Peterson
> http://braverock.com/brian/
> Ph: 773-459-4973
> IM: bgpbraverock
> _______________________________________________
> R-sig-hpc mailing list
> R-sig-hpc at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-hpc


blindglobe at gmail.com
Muttenz, Switzerland.
"Commit early,commit often, and commit in a repository from which we
can easily roll-back your mistakes" (AJR, 4Jan05).

Drink Coffee:  Do stupid things faster with more energy!

More information about the R-sig-hpc mailing list