Bureaucrats, cc_docs_admin, cc_staff
2,773
edits
No edit summary |
(tweak LAPACK BLAS -> MKL instructions) |
||
Line 47: | Line 47: | ||
== BLAS/LAPACK and MKL == <!--T:10--> | == BLAS/LAPACK and MKL == <!--T:10--> | ||
Many software packages expect the commonly used numerical linear algebra libraries BLAS and LAPACK to be available as <tt>-lblas</tt> and <tt>-llapack</tt>. This is not the case on Compute Canada systems, where these libraries are integrated with Intel's [https://software.intel.com/en-us/mkl Math Kernel Library] (MKL). If you are using one of the Intel compilers (e.g. ifort, icc | Many software packages expect the commonly used numerical linear algebra libraries BLAS and LAPACK to be available as <tt>-lblas</tt> and <tt>-llapack</tt>. This is not the case on Compute Canada systems, where these libraries are integrated with Intel's [https://software.intel.com/en-us/mkl Math Kernel Library] (MKL). If you are using one of the Intel compilers (e.g. <code>ifort, icc, icpc</code>) then the solution is to replace <tt>-lblas</tt> and <tt>-llapack</tt> with <tt>-mkl=sequential</tt> (without internal MKL threading) or <tt>-mkl</tt> (with threading) in your compiler and linker options in order to ensure that the MKL and thus BLAS/LAPACK are used. See [https://software.intel.com/en-us/mkl-linux-developer-guide-using-the-mkl-compiler-option here] for more on the significance of <code>sequential</code> and other options. | ||
<!--T:23--> | <!--T:23--> |