Automatically tuned linear algebra software wiki




















The FSF is a charity with a worldwide mission to advance software freedom — learn about our history and work. Privacy Policy. JavaScript license information. Read more. Free Software Foundation! From Free Software Directory. Jump to: navigation , search. Licensing License.

Choosing the bigger of M or N for the outer loop reduces the footprint of the copy. Instead it defines a parameter, Kp , to give best use of the L2 cache. Panels are limited to Kp in length. If that fails it uses the no-copy version of GEMM, but this case is unlikely for reasonable choices of cache edge. Kp is a function of cache edge and NB. Installing ATLAS on a particular platform is a challenging process which is typically done by a system vendor or a local expert and made available to a wider audience.

For many systems, architectural default parameters are available; these are essentially saved searches plus the results of hand tuning.

On such systems the installation process is greatly simplified. Copyright The image is from Wikipedia Commons. Wikipedia Page. Also, users may want to choose a particular implementation that works best for them at run time. This guideline gives a structure that can enforce the first while allowing the second, as well as providing a transparent fallback mechanism to Netlib's reference implementation for those symbols not included in the selected backend. It also provides transparent fallback to Netlib's reference implementation if a certain symbol is not present in the selected backend.

These are the main features:. These are sub-packaged separately to avoid pulling the whole stack of implementations. The netlib and netlib64 sub-packages require a default implementation that is currently set to openblas-serial:.



0コメント

  • 1000 / 1000