Applies ToRevolution Analytics

Revolution R ships with optimized Intel MKL Blas libraries and a different version of Lapack than is used with CRAN-R by default. For most computations you can do in R these will be faster than the default ones shipped with CRAN-R. However, this is not the case for all computations. As well, some third-party R packages may not be compatible with the MKL Blas libraries. In these cases, you may want to revert to the default libraries shipped with CRAN-R.Below are examples on how to perform this swap for Revolution R Enterprise 7.1. Note that the CRAN-R libraries must come from the same level of R as that shipped with Revolution R. For versions 7.1, this is R 3.0.2.Windows:

  1. Rename 'Rblas.dll' to 'Rblas.dll.RevoR' and 'Rlapack.dll' to 'Rlapack.dll.RevoR' in the folder  'C:\Revolution\R-Enterprise-7.1\R-3.0.2\bin\x64'.

  2. Copy Rblas.dll and Rlapack.dll from the <R_HOME>\lib directory of a CRAN-R 3.0.2 installation to C:\Revolution\R-Enterprise-7.1\R-3.0.2\bin\x64.

  3. Restart Revolution R Enterprise, for the changes to take effect.

Linux:

  1. Rename libRblas.so and libRlapack.so in /usr/lib64/Revo-7.1/R-3.0.2/lib64/R/lib:

    # mv /usr/lib64/Revo-7.1/R-3.0.2/lib64/R/lib/libRblas.so /usr/lib64/Revo-7.1/R-3.0.2/lib64/R/lib/libRblas.so.RevoR# mv /usr/lib64/Revo-7.1/R-3.0.2/lib64/R/lib/libRlapack.so /usr/lib64/Revo-7.1/R-3.0.2/lib64/R/lib/libRlapack.so.RevoR
  2. Copy libRblas.so and libRlapack.so from the <R_HOME>/lib directory of a CRAN-R 3.0.2 installation to /usr/lib64/Revo-7.1/R-3.0.2/lib64/R/lib.

  3. Restart Revolution R Enterprise, for the changes to take effect.

Need more help?

Want more options?

Explore subscription benefits, browse training courses, learn how to secure your device, and more.

Communities help you ask and answer questions, give feedback, and hear from experts with rich knowledge.