Draft: zgetrf : MPI all reduce in permutation

changed milestone to %Chameleon 1.3.0

added StarPU label

requested review from @faverge

assigned to @alisito

added 22 commits

• 71136b35...06e70618 - 11 commits from branch solverstack:master
• 828ad37a - 1 earlier commit
• 00b1f0ef - starpu/codelet/gemm: code indent
• e9e7258b - zgetrf batched: Add an a field in the data structure and environment variable...
• 45a51845 - zgetrf batched: Add an a cmake option to set the maximum batch size
• f2fedccb - zgetrf batched: create codelet_zgetrf_batched.c
• 19cd7487 - zgetrf batched: add codelet batched for getrf percol
• 9c4ab968 - zgetrf batched: add batched percol algorithm
• d16d051c - zgetrf batched: add codelet batched for getrf blocked
• 12d3bac8 - zgetrf batched: add batched blocked algorithm
• 6280fa6f - zlaswp batched: add batched zlaswp codelet
• 7c74ff67 - zlaswp batched: add batched zlaswp algorithm

Compare with previous version

added 4 commits

• f710776d - zgetrf batched: add codelet batched for getrf blocked
• 3d2245f2 - zgetrf batched: add batched blocked algorithm
• 0d4313f9 - zlaswp batched: add batched zlaswp codelet
• aba893f0 - zlaswp batched: add batched zlaswp algorithm

Compare with previous version

added 2 commits

• 38d04999 - zlaswp batched: add batched zlaswp codelet
• cdac8e44 - zlaswp batched: add batched zlaswp algorithm

Compare with previous version

added 129 commits

• cdac8e44...3ad53bc1 - 113 commits from branch solverstack:master
• 3ad53bc1...01326a02 - 6 earlier commits
• 3d753a4a - zgetrf: update CTestLists.cmake
• 02e1d30a - zgetrf: add quark, parsec and openmp zipiv_redux codelet and update runtime_descriptor_ipiv
• 190d60a8 - zgetrf: correct debug cppi print
• 5dcf9f52 - zgetrf: Add all_reduce algorithm choice
• eb0ebf8d - zgetrf: StarPU all_reduce without tasks algorithm
• 7f05c594 - zgetrf: StarPU MPI all_reduce algorithm
• 354fda4e - zgetrf: MPI only all_reduce algorithm
• bcd3b207 - zgetrf: StarPU MPI all_reduce in task algorithm
• 205a347e - zlaswp batched: add batched zlaswp codelet
• 0dac60ee - zlaswp batched: add batched zlaswp algorithm

Compare with previous version

added 2 commits

• 1895efdf - compute/pzgetrf.c
• d839cedd - runtime/starpu/codelets/codelet_zlaswp_batched.c

Compare with previous version

added 13 commits

• d839cedd...c2d5b5b7 - 3 earlier commits
• c1fcd9ac - zgetrf: correct debug cppi print
• 52a80f45 - zgetrf: Add all_reduce algorithm choice
• ff570bfb - zgetrf: StarPU all_reduce without tasks algorithm
• 6e25b332 - zgetrf: StarPU MPI all_reduce algorithm
• a47b8ea1 - zgetrf: MPI only all_reduce algorithm
• ea3ff244 - zgetrf: StarPU MPI all_reduce in task algorithm
• 115d7f34 - zlaswp batched: add batched zlaswp codelet
• 49a32a2e - zlaswp batched: add batched zlaswp algorithm
• 9b93bdd6 - compute/pzgetrf.c
• 3281fb63 - runtime/starpu/codelets/codelet_zlaswp_batched.c

Compare with previous version

added 13 commits

• 3281fb63...a9d74b76 - 3 earlier commits
• e154d05c - zgetrf: correct debug cppi print
• 9a9340ca - zgetrf: Add all_reduce algorithm choice
• 0feac445 - zgetrf: StarPU all_reduce without tasks algorithm
• bd7f5b30 - zgetrf: StarPU MPI all_reduce algorithm
• 91e41852 - zgetrf: MPI only all_reduce algorithm
• 885a9dd5 - zgetrf: StarPU MPI all_reduce in task algorithm
• 8aeab729 - zlaswp batched: add batched zlaswp codelet
• 3bbd23d5 - zlaswp batched: add batched zlaswp algorithm
• f01fed8d - compute/pzgetrf.c
• 532f800b - runtime/starpu/codelets/codelet_zlaswp_batched.c

Compare with previous version

added 12 commits

• 532f800b...f63859df - 2 earlier commits
• d00234bd - zgetrf: correct debug cppi print
• cc06a228 - zgetrf: Add all_reduce algorithm choice
• f921b21c - zgetrf: StarPU all_reduce without tasks algorithm
• 405117f2 - zgetrf: StarPU MPI all_reduce algorithm
• 508b36b6 - zgetrf: MPI only all_reduce algorithm
• 3a0a1ee2 - zgetrf: StarPU MPI all_reduce in task algorithm
• cf859bdb - zlaswp batched: add batched zlaswp codelet
• 3d45da34 - zlaswp batched: add batched zlaswp algorithm
• 1cbbd8df - compute/pzgetrf.c
• 6bafc8b9 - runtime/starpu/codelets/codelet_zlaswp_batched.c

Compare with previous version

added 25 commits

• 6bafc8b9...790f4633 - 13 commits from branch solverstack:master
• 790f4633...a53772cb - 2 earlier commits
• 1b218fef - zgetrf: Add all_reduce algorithm choice
• dfbbc15b - zgetrf: StarPU all_reduce without tasks algorithm
• 9f191b88 - zgetrf: StarPU MPI all_reduce algorithm
• d88114e0 - zgetrf: MPI only all_reduce algorithm
• 3db2acdf - zgetrf: StarPU MPI all_reduce in task algorithm
• df825273 - Dedicated task to print stats by step. Execute task even with SimGrid support.
• 7bbac941 - zlaswp batched: add batched zlaswp codelet
• aee0dbb0 - zlaswp batched: add batched zlaswp algorithm
• 41f75c7b - compute/pzgetrf.c
• f4b42f14 - runtime/starpu/codelets/codelet_zlaswp_batched.c

Compare with previous version

added 1 commit

• f6cb8a82 - getrf: function to get the array of proc involved in panel

Compare with previous version

added 1 commit

• 3b661f9b - getrf: function to get the array of proc involved in panel

Compare with previous version

added 19 commits

• 3b661f9b...4dcb5308 - 9 earlier commits
• 79037741 - zlaswp batched: add batched zlaswp algorithm
• 8ef3c623 - compute/pzgetrf.c
• 92a24243 - runtime/starpu/codelets/codelet_zlaswp_batched.c
• 138a648f - getrf: function to get the array of proc involved in panel
• ba9ad32b - compute/pzgetrf.c
• 0581c397 - runtime/starpu/codelets/codelet_zperm_allreduce.c
• 37e3491a - runtime/starpu/codelets/codelet_zipiv_allreduce.c
• 0cbbc591 - runtime/starpu/codelets/codelet_zlaswp.c
• 3af58beb - control/compute_z.h
• 16aba3c3 - compute/zgetrf.c

Compare with previous version

added 5 commits

• 0ae29388 - runtime/starpu/codelets/codelet_zperm_allreduce.c
• 47b0445a - runtime/starpu/codelets/codelet_zipiv_allreduce.c
• 0194fda5 - runtime/starpu/codelets/codelet_zlaswp.c
• e796f419 - control/compute_z.h
• c1d9a36b - compute/pzgetrf.c

Compare with previous version

added 18 commits

• c1d9a36b...0963fcd9 - 8 earlier commits
• 5708afa9 - zlaswp batched: add batched zlaswp codelet
• 3a99e1ec - zlaswp batched: add batched zlaswp algorithm
• e725970a - compute/pzgetrf.c
• c9893dae - runtime/starpu/codelets/codelet_zlaswp_batched.c
• 68042428 - getrf: function to get the array of proc involved in panel
• ffd3ba7d - runtime/starpu/codelets/codelet_zperm_allreduce.c
• 6a79c7a4 - runtime/starpu/codelets/codelet_zipiv_allreduce.c
• 50d2ac5a - runtime/starpu/codelets/codelet_zlaswp.c
• 40d03d4e - control/compute_z.h
• a1c55cb4 - compute/pzgetrf.c

Compare with previous version

added 6 commits

• d0068199 - runtime/starpu/codelets/codelet_zipiv_allreduce.c
• 86f4b58a - zlaswp batched: add batched zlaswp codelet
• 91f9335a - zlaswp batched: add batched zlaswp algorithm
• 2115a58f - zgetrf: function to get the array of proc involved in panel
• b9b90b21 - zgetrf: add codelet zperm allreduce
• 7a362292 - zgetrf: zperm all reduce algorithm

Compare with previous version

added 8 commits

• faa4dd8f - zgetrf: MPI only all_reduce algorithm
• f1d20b38 - zgetrf: StarPU MPI all_reduce in task algorithm
• b824ab85 - Dedicated task to print stats by step. Execute task even with SimGrid support.
• 2412c71f - zlaswp batched: add batched zlaswp codelet
• 6af5b1f4 - zlaswp batched: add batched zlaswp algorithm
• ce40cead - zgetrf: function to get the array of proc involved in panel
• 5d2a85d8 - zgetrf: add codelet zperm allreduce
• 009085d2 - zgetrf: zperm all reduce algorithm

Compare with previous version