Added _starpu_mpi_redux_wrapup_data detection so that automatic MPI reduction can be used for partial pivoting (see also RUNTIME_ipiv_reducek).

starpu/lacpy: Fix a huge memory leak whan using the starpu_data_cpy family functions instead of lacpy

starpu/lacpy: make sure both src and dst tiles have the same size when using the starpu copy mechanism

PRUVOST Florent authored 1 year ago and Mathieu Faverge committed 1 year ago

Fix compilation problem with starpu versions < 1.4. STARPU_CUDA and STARPU_HIP macros may not be defined in starpu

Fix --forcegpu option to be able to force starpu to execute kernels on gpus, RUNTIME_CUDA was not well defined and introduce RUNTIME_HIP.

Matthieu Kuhn authored 1 year ago and Mathieu Faverge committed 1 year ago

runtime: Add struct fields and functions to manipulate the permutation and the inverse permutation arrays

Added permutation on top of partial pivoting, panel only.
First version of LU with pivoting.

codelet/getrf_percol: Add a new codelet_zgetrf_percol file with the codelets for GETRF with partial pivoting per column

runtime/starpu: extract the wontuse and mpi_data_register functions into header files to share them into multiple C files

runtime/starpu/control/runtime_descriptor.c

Matthieu Kuhn authored 1 year ago and Mathieu Faverge committed 1 year ago

starpu/codelet/getrf_nopiv: move nopiv percol codelets to a dedicated file and rename them (remove panel)