Package Details: r-mkl 4.4.1-1

Git Clone URL: https://aur.archlinux.org/r-mkl.git (read-only, click to copy)
Package Base: r-mkl
Description: Language and environment for statistical computing and graphics, linked to the Intel(R) MKL.
Upstream URL: http://www.r-project.org/
Keywords: hpc mathematics modelling r statistics
Licenses: GPL
Conflicts: microsoft-r-open, r
Provides: r
Submitter: giniu
Maintainer: alexanderp
Last Packager: alexanderp
Votes: 25
Popularity: 0.000108
First Submitted: 2010-05-06 00:10 (UTC)
Last Updated: 2024-07-04 19:34 (UTC)

Required by (3400)

Sources (5)

Latest Comments

« First ‹ Previous 1 .. 15 16 17 18 19 20 21 Next › Last »

jdarch commented on 2013-10-24 12:15 (UTC)

aberkoke: that's right, the icu-package (icu: International Components for Unicode library) has moved from version 51.x to 52.x 12 days ago. You had built r-mkl with version 51.x, so now I cannot find the library anymore, as R is looking for libicu*.so.51, but the libs are now called libicu*.so.52. Solution: build r-mkl again and it will work again.

aberkoke commented on 2013-10-24 10:50 (UTC)

R doesn't work after latest upgrades. There is a problem with package "icu" or "lib32-icu". [eduardo@manjaro-linux lib]$ R /usr/lib64/R/bin/exec/R: error while loading shared libraries: libicuuc.so.51: cannot open shared object file: No such file or directory

jdarch commented on 2013-10-22 02:31 (UTC)

To speed up some more commonly used basic math functions (e.g. log,root,trigonometric) I have tested linking against Intel's 'libimf' (included in intel-mkl) and AMD's 'libamdlibm' (in AUR: amdlibm). Use of amdlibm resulted in substantial speedups compared to libm-218, libimf also showed speedups, but much less substantial (at least on the (AMD) CPU I had used for testing). For both libraries some functions showed slight slowdowns compared to libm as well. It would be great to get some more input from users with other processors to decide if linking against libimf or libamdlibm by default makes sense. You can try out the libraries by preloading them (LD_PRELOAD).

jdarch commented on 2013-10-04 07:23 (UTC)

aberkoke, I have never compared the performance of R compiled with Intel's compilers to R compiled with the GNU compiler collection. I would not expect a substantial difference in most cases, but that might vary dependant on the type of CPU and what instructions R needs to perform. But if you could try it out, that would be great! It should be possible to make different versions with different compilers and compiler flags by playing with the CC, CFLAGS etc. environment variables before building the package.

aberkoke commented on 2013-10-03 12:11 (UTC)

Thanks for the maintenance of the pagacke! What do you think about compile R with intel icc? better performance?

jdarch commented on 2013-09-25 10:22 (UTC)

Updated to 3.0.2. If there are file conflicts because of updated packages, just delete the 3.0.1 files. Please inform me of any bugs.