Releases: RevolutionAnalytics/RRO
RRO-8.0.3
This release of Revolution R Open is based on R-3.1.3 and adds the performance enhancements from the Intel MKL (optional additional install).
Release Highlights
- Based on the R-3.1.3 engine
- Adds support for Ubuntu 15.04
- Uses a static MRAN snapshot at 2015-04-01 as the CRAN mirror
- Installs
checkpoint0.3.10 from MRAN for package reproducibility - On Windows and Linux platforms, use of the Intel MKL is now optional. Download and instal the MKL Math Library installer separately to enable multi-threaded math on these platforms. (On Mac, the Accelerate Framework is always used.)
Fixed Bugs
- Fixes the instability problems some users encountered when using the Intel MKL with RRO-8.0.2 on Windows and RStudio
- Fixes an issue with running packages that use tcl/tk within the RRO GUI (RRO.app) on OS X [Issue #107]
Known Issues
devtools::install_githubmay fail with the error "error in running command" on Ubuntu. To work around this issue, setoptions(unzip = 'internal')before callinginstall_githib. (See #37 and r-lib/devtools#406 )
Binary Downloads
Visit http://mran.revolutionanalytics.com/download/ to download the binary distribution.
RRO-8.0.2-beta
This release of Revolution R Open fixes some bugs, and uses the same R engine as RRO 8.0.1. We recommend upgrading to this version if you are experiencing any of the issues listed in the “Fixed bugs” section.
Highlights
- Based on R-3.1.2
- On Windows and Linux platforms, use of the Intel MKL is now optional. Download the MKL Math Library installer separately and install to enable multi-threaded math on these platforms. (On Mac, the Accelerate Framework is always used.)
Fixed bugs
An issue with missing library files (libRlapack.dylib) on Mac has been resolved (#88)
Known issues
The RRO GUI (RRO.app) on OS X fails whenever you attempt to load a package that launches a tcltk window (for example, the HH package). To work around this issue, use R in a terminal window or another IDE such as RStudio for these packages. [Issue #107]
devtools::install_github may fail with the error "error in running command" on Ubuntu. To work around this issue, set options(unzip = 'internal') before calling install_githib. (See #37 and r-lib/devtools#406 )
With the MKL libraries installed, many commands cause a crash when using RRO within RStudio on Windows. To workaround this issue, you can replace your Rprofile.site file as follows: (Issue #119)
- Create a back up copy of your
Rprofile.sitefile found here:C:\Program Files\RRO\R-3.1.2\etc\Rprofile.site. - Download an updated copy of
Rprofile.site. - Save the downloaded copy of
Rprofile.siteto the directoryC:\Program Files\RRO\R-3.1.2\etc. - In the Copy File dialog, choose Copy and Replace.
- In the Destination Folder Access Denied dialog, click Continue.
- Restart RStudio.
Download the binaries
Visit http://mran.revolutionanalytics.com/download/ to download the binary distribution.
RRO-8.0.1-beta3
Released only on Ubuntu platforms.
Updated Ubuntu builds to include tcl/tk.
Known issues
- After installing R on top of Revolution R Open, the Revolution R Open banner message may still appear when starting R. (Uninstall RRO before installing R to avoid this problem.)
- BioConductor fails to install on Ubuntu. This workaround will fix the problem:
cd /usr/lib64/RRO-8.0.1/R-3.1.2/lib/R/etc/
wget https://raw.githubusercontent.com/RevolutionAnalytics/RRO/master/R-src/etc/repositories
Now the following commands in RRO will successfully install BioConductor:
source("http://bioconductor.org/biocLite.R")
biocLite()Visit http://mran.revolutionanalytics.com/download/ to download the binary distribution.
RRO-8.0.1-beta2
Released only for Mac OS X platforms.
Mac build now uses Mac Accelerate Framework instead of Intel MKL.
Known issues
- After installing R on top of Revolution R Open, the Revolution R Open banner message may still appear when starting R. (Uninstall RRO before installing R to avoid this problem.)
- On Mac, Some packages (e.g. igraph) may complain about a missing libRlapack.dylib . (#88)
If you attempt to load an R package and get an error message stating that the BLAS and/or LAPACK math library is missing, then you will have to manually download, copy and link to that library. Here’s how:- Download the BLAS library and the LAPACK library to your file system.
- Copy the libraries using these commands:
sudo cp libRblas.0.dylib /Library/Frameworks/R.framework/Libraries
sudo cp libRlapack.dylib /Library/Frameworks/R.framework/Libraries - Link the BLAS library using this command:
sudo ln -s /Library/Frameworks/R.framework/Libraries/libRblas.0.dylib /Library/Frameworks/R.framework/Libraries/libRblas.dylib
Visit http://mran.revolutionanalytics.com/download/ to download the binary distribution.
RRO-8.0.1-beta
Known issues
- Revolution R Open.app uses Mac Accelerators for performance, instead of the intended MKL. (The command-line application R uses MKL as intended. Using RRO from RStudio also uses the MKL.) For this reason, the Revobase package isn't available when running Revolution R Open.app. (fixed in v8.0.1-beta2)
- The installer is large on Mac. (This is because two R frameworks are installed: R.framework which uses Accelerators; and RRO.framework, which uses MKL.) (fixed in v8.0.1-beta2)
- After installing R on top of Revolution R Open, the Revolution R Open banner message may still appear when starting R. (Uninstall RRO before installing R to avoid this problem.)
- tcl/tk support is lacking in the Ubuntu build (#79) (fixed in v8.0.1-beta3)
- BioConductor fails to install on CentOS. This workaround will fix the problem:
cd /usr/lib64/RRO-8.0.1/R-3.1.2/lib64/R/etc/
wget https://raw.githubusercontent.com/RevolutionAnalytics/RRO/master/R-src/etc/repositories
Now the following commands in RRO will successfully install BioConductor:
source("http://bioconductor.org/biocLite.R")
biocLite()Visit http://mran.revolutionanalytics.com/download/ to download the binary distribution.
8.0.0 beta
Visit http://mran.revolutionanalytics.com/archives/ to download the binary distribution.