Sorry for the trouble with double posting: I never received my original message 
through the mailing list so I assumed that it did not go through.

I’ll try your suggestions, thank you.

Mikhail.

> On Nov 14, 2014, at 12:04, Jeff Newmiller <jdnew...@dcn.davis.ca.us> wrote:
> 
> Repeating yourself verbatim is poor form. At least quote your previous 
> message and indicate that this is a reprise.
> 
> While I doubt that my opinion is universal, I think that mixing C and R in 
> the same file is already living on the edge of good practice. You are 
> assuming that all of your nodes have the infrastructure to compile your code 
> on the fly, which is something you have to insure for each node in your 
> facilities. A less convenient but more conventional approach is to create 
> your own support package(s) that contain the compiled portions of your code 
> and distribute them prior to running your application.
> 
> However, this conversation might be more productive if conducted between you 
> and the maintainer of the inline package (see ?maintainer) or on the R-devel 
> list because it is outside the R language.
> 
> ---------------------------------------------------------------------------
> Jeff Newmiller                        The     .....       .....  Go Live...
> DCN:<jdnew...@dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
>                                      Live:   OO#.. Dead: OO#..  Playing
> Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
> /Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
> --------------------------------------------------------------------------- 
> Sent from my phone. Please excuse my brevity.
> 
> On November 14, 2014 7:42:01 AM PST, Mikhail Umorin <mike...@gmail.com> wrote:
>> Hello —
>> 
>> I am using inline C functions within foreach %dopar% loop. On SMP
>> (doParallel, doMP) it works but on MPI-based clusters (doMPI) it
>> does’t. The reason, I think, is because the object code produced using
>> the inline package, essentially an .so file, does not get copied onto
>> the worker nodes when they are spawned using startMPIcluster().
>> 
>> Does anyone have any ideas on how I can send the object code to the
>> workers and re-link it to the inline functions, or some other way? I
>> know that I can define the inline functions within the foreach loop but
>> it is very limiting and does not look pretty when you have a lot of
>> functions to define. So, is there I way that follows a good programming
>> style?
>> 
>> I would prefer to do it from within the R script and use the “spawning”
>> mode to paralellize.
>> 
>> Thank you for your time, 
>> 
>> Mikhail.
>> ______________________________________________
>> 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.
> 

______________________________________________
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.

Reply via email to