rsnt_translations
56,430
edits
BLAS and LAPACK: Difference between revisions
Jump to navigation
Jump to search
no edit summary
No edit summary |
No edit summary |
||
| Line 11: | Line 11: | ||
= Which implementation should I use ? = <!--T:4--> | = Which implementation should I use ? = <!--T:4--> | ||
For the past few years, we have been recommending to use Intel MKL as a reference implementation. This recommendation was driven by the fact that we only had Intel processors in our clusters. This changed with the arrival of Narval, which is built with AMD processors. We now recommend using FlexiBLAS when compiling code. Our FlexiBLAS module is configured such that Intel MKL will be used except when using AMD processors, in which case BLIS will be used. This arrangement will usually offer optimal performance. | For the past few years, we have been recommending to use Intel MKL as a reference implementation. This recommendation was driven by the fact that we only had Intel processors in our clusters. This changed with the arrival of [[Narval/en|Narval]], which is built with AMD processors. We now recommend using FlexiBLAS when compiling code. Our FlexiBLAS module is configured such that Intel MKL will be used except when using AMD processors, in which case BLIS will be used. This arrangement will usually offer optimal performance. | ||
= How do I compile against FlexiBLAS ? = <!--T:5--> | = How do I compile against FlexiBLAS ? = <!--T:5--> | ||