Commits · master · Yifan Yao / inq · GitLab

Jun 19, 2021
- Merge branch 'update_BMPI3' into 'master' · 85cd9d0f
  Alfredo Correa authored Jun 19, 2021
  85cd9d0f
- update bmpi3 · f192167b
  Alfredo Correa authored Jun 18, 2021
  
  f192167b
- Merge branch 'alltoall_optimization' into 'master' · d96f935a
  Xavier Andrade authored Jun 19, 2021
  d96f935a
- update bmpi3 · befb586c
  Alfredo Correa authored Jun 18, 2021
  
  befb586c
Jun 18, 2021
- Merge branch 'correaa-master-patch-04257' into 'master' · 604c5c76
  Alfredo Correa authored Jun 18, 2021
  604c5c76
- Merge branch 'remove_comm_copy_contructor' into 'master' · 3a11acbb
  Alfredo Correa authored Jun 18, 2021
  3a11acbb
Jun 17, 2021
- Update README.md · ee27bf9c
  Alfredo Correa authored Jun 17, 2021
  
  ee27bf9c
- Do alltoall using point to point communication. · 2d5a7d4a
  Xavier Andrade authored Jun 17, 2021
  2d5a7d4a
- Make gpu::alltoall receive an array. · b1e4854c
  Xavier Andrade authored Jun 17, 2021
  b1e4854c
- Deduct the MPI type from the pointer. · b0e5dadb
  Xavier Andrade authored Jun 17, 2021
  
  b0e5dadb
- Use the wrapper in the FFTs. · b78c55fb
  Xavier Andrade authored Jun 17, 2021
  
  b78c55fb
- Removed redundant arguments. · 15579e8c
  Xavier Andrade authored Jun 17, 2021
  
  15579e8c
- Added gpu::alltoall, a wrapper for MPI_Alltoall. · 6cb28e81
  Xavier Andrade authored Jun 17, 2021
  
  6cb28e81
- prepare inq for non copiable mpi3 communicator · bfc5abf4
  Alfredo Correa authored Jun 16, 2021
  
  bfc5abf4
Jun 16, 2021
- take communicator by move · 643bbbdf
  Alfredo Correa authored Jun 16, 2021
  
  643bbbdf
- Merge branch 'parallel_optimiztion' into 'master' · c45aa78c
  Xavier Andrade authored Jun 16, 2021
  c45aa78c
- Do not use multi for the transposition in the parallel FFT. · 0c9b95ce
  Xavier Andrade authored Jun 15, 2021
  
  0c9b95ce
- Merge branch 'fix_broyden_parallel' into 'master' · 88c23f2c
  Xavier Andrade authored Jun 16, 2021
  88c23f2c
Jun 15, 2021
- Merged two kernels in one. · 51af5402
  Xavier Andrade authored Jun 15, 2021
  
  51af5402
- Bugfix: the parallelization of Broyden was incorrect. · 4c725ec7
  Xavier Andrade authored Jun 15, 2021
  
  4c725ec7
- Merge branch 'correaa-master-patch-58968' into 'master' · 4eee1778
  Xavier Andrade authored Jun 15, 2021
  4eee1778
Jun 10, 2021
- Update README.md · 4da7b4cc
  Alfredo Correa authored Jun 10, 2021
  
  4da7b4cc
- Merge branch 'do_not_calculate_forces' into 'master' · b60388d4
  Xavier Andrade authored Jun 10, 2021
  View commits for tag v0.5.0 v0.5.0
  
  b60388d4
- Merge branch 'batched_gemms_projector' into 'master' · 5f7448b3
  Xavier Andrade authored Jun 10, 2021
  5f7448b3
- Merge branch 'parallel_fixes' into 'master' · 71cb820d
  Xavier Andrade authored Jun 10, 2021
  71cb820d
Jun 09, 2021
- Do not calculate the ground state forces by default. · bfe618b2
  Xavier Andrade authored Jun 09, 2021
  
  bfe618b2
- Do not calculate the forces at the beggining of TD, unless needed for · 95d2bfad
  Xavier Andrade authored Jun 09, 2021
  95d2bfad
- Removed some left over debugging stuff. · acd61448
  Xavier Andrade authored Jun 09, 2021
  
  acd61448
- Use a batched gemm in the second half of the projection. · a49d0b53
  Xavier Andrade authored Jun 09, 2021
  
  a49d0b53
- Sync after cublas. · 9c78d9fa
  Xavier Andrade authored Jun 09, 2021
  
  9c78d9fa
- Avoid the size 0 case. · d36ec4ec
  Xavier Andrade authored Jun 09, 2021
  
  d36ec4ec
- Use batched gemm for the non local projector. · 6670aee8
  Xavier Andrade authored Jun 09, 2021
  
  6670aee8
- Add a profiling region for the communication in diagonalize. · 1c349df1
  Xavier Andrade authored Jun 08, 2021
  
  1c349df1
- Do the dense diagonalization in one node and broadcast. · 7e988098
  Xavier Andrade authored Jun 08, 2021
  
  7e988098
- Cannot use the collective version of file reads here. · 880efeae
  Xavier Andrade authored Jun 08, 2021
  
  880efeae
Jun 08, 2021
- Merge branch 'fix_parallelization_topology' into 'master' · d71ecf16
  Xavier Andrade authored Jun 08, 2021
  d71ecf16
- Merge branch 'aluminum_supercell' into 'master' · 621a0ac7
  Xavier Andrade authored Jun 08, 2021
  621a0ac7
- Select the paralelization for the real time test. · 92261200
  Xavier Andrade authored Jun 07, 2021
  
  92261200
- Allow the gs and td calculations to use a different parallelization. · 0dbe9277
  Xavier Andrade authored Jun 07, 2021
  
  0dbe9277
- Do a TD calculation in aluminum. · 02bdd381
  Xavier Andrade authored Jun 07, 2021
  
  02bdd381