]]>

Valid question. Answer: It may just be a matter of taste.

I initially went for the (complicated & headless) installation from source since I wasn't sure if using the Epiphany processor from R would require changes to the R source itself. But that is not the case, since we can go the OpenCL route.

S.

Statistics: Posted by censix — Thu Jun 05, 2014 9:58 pm

]]>

bit new in using R, i tried installing using the headless one and succeeded but then i realized that i needed the graphical interface so what i did is removed the headless R by using make uninstall and reinstalled by means of sudo apt-get install r-base r-base-dev

what is the advantage of using the headless installation compared to the basic installation procedure?

thanks appreciate your help.

kind regards,

Jubert

Statistics: Posted by jubertroldan — Thu Jun 05, 2014 12:39 pm

]]>

Statistics: Posted by shodruk — Wed May 28, 2014 3:29 pm

]]>

actually, there is a quick and dirty way to get R running on the parallella board, by simply installing r-cran-* and other binary packages directly from the linaro package repository. At the command prompt on you parallella board(s), do:

# for the prerequisites

sudo apt-get install gfortran libreadline6-dev

# then, you could do

sudo apt-get install r-cran-rjava

# which will install all the dependencies needed for running R

It has been working ok for me.

If you use this method, you should also automatically have X11 support, i.e. graphical output from R.

Soren

http://censix.com

Statistics: Posted by censix — Tue May 27, 2014 8:59 pm

]]>

One question: will R still work with parallela if I use the x11 interface, or do I need to learn to do everything command line? Is the best choice for someone like me putting the R programs together on another machine and just copying them over to run them (or waiting until someone else does all the hard work of making it more user friendly)? It still takes me tons of work to figure out how to run new things from the command line, but I do ok with learning a couple of commands.

Thanks for any advice.

Statistics: Posted by theyogre — Tue May 06, 2014 9:03 pm

]]>

all that is needed is a BLAS library that uses the epiphany, i.e. through OpenCL. In another post I have suggested that we could use the "viennacl" for this.

Some clarifications:

This BLAS library does in no way have to be specific to R !

In addition, with a few additional ./configure flags, R can be compiled in a way so that it uses an external (epiphany driven) BLAS library. In that scenario, again, R does not care how the BLAS library does the calcuations.

For comparison:

1) Lets say you have an nvidia graphics card and install the cuda/opencl capable dirvers for it. These drivers come with a library called "cuBLAS" which implements BLAS routines using the nvidia GPU though cuda.

2) Lets say you have an ati graphics card and install the opencl drivers. Then again, these dirvers come with a library called "libclamdblas" that implements BLAS routines using the ati GPU.

In both scenarios, 1) and 2) it is possible to configure R so that it will use 'cuBLAS' or 'libclamdblas' instead of its own built in BLAS library, therby using the GPUs to accelerate almost all basic R calculations!

So coming back to the original point. If someone manages to create a standard BLAS implementation of as many levels as possible (1-3) using the epiphany processor ... we have won the game! Like I said. Viennacl look like a very promising starting point.

Statistics: Posted by censix — Fri Jun 28, 2013 8:49 pm

]]>

There is some hope though.

Tim

Statistics: Posted by timpart — Fri Jun 28, 2013 6:37 pm

]]>

]]>

ysapir wrote:

Do you know where in the R's source would the math library be, and how to interface with it (in terms of required data structures, etc.)?

Do you know where in the R's source would the math library be, and how to interface with it (in terms of required data structures, etc.)?

If you are referring to the the BLAS (Basic Linear Algebra Subprograms) API which can switched in R for a different implementation there is a specification here http://www.netlib.org/blas/blast-forum/ and more information here http://www.netlib.org/blas/

Looks promising for acceleration, especially the higher levels which involve more calculation.

Tim

Statistics: Posted by timpart — Thu Jun 27, 2013 6:30 am

]]>

However, R is a high memory demanding environment, especially when dealing with big chunk of data or large simulation. Do you have any suggestion on how to overcome this?

I wonder if in the future Parallella will support extendable memory banks?

Cheers

Statistics: Posted by juanlp — Thu Jun 27, 2013 2:50 am

]]>

good to know this is useful to others. It certainly is to me.

I personally dont have any benchmark numbers for GPU accelerated R, but am sure they are out there somewhere.

Octave yes ... or Python...we just need to find an appropiate Guru.

Soren

http://censix.com

Statistics: Posted by censix — Thu Jun 13, 2013 9:14 pm

]]>

Andreas

Statistics: Posted by aolofsson — Thu Jun 13, 2013 12:56 pm

]]>

I would look into the R source tree somewhere here

R-2.15.1/src/modules/lapack/

The configure flags needed to use an external BLAS are probably something like

./configure --enable-BLAS-shlib --enable-R-shlib .....

(taken from this nice document)

http://www.rochester.edu/college/gradst ... g/BLAS.pdf

Statistics: Posted by censix — Thu Jun 13, 2013 9:35 am

]]>

Do you know where in the R's source would the math library be, and how to interface with it (in terms of required data structures, etc.)?

Statistics: Posted by ysapir — Wed Jun 12, 2013 8:27 pm

]]>