QR Factorization Using Malleable BLAS on Multicore Processors.

We demonstrate that significant performance benefits can be obtained via the exploitation of malleability in a framework designed to implement portable and high-performance BLAS-like kernels. For this purpose, we integrate thread-level malleability within the BLIS library, providing an experimental evaluation for a representative dense linear algebra operation such as the QR factorization for dense matrices enhanced with look-ahead.