Skip to content

Actions: NVIDIA/cuDecomp

Actions

Build

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
build.yml
94 workflow runs
94 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
Use native NVSHMEM synchronization APIs in NVSHMEM backends Build #70: Issue comment #107 (comment) created by romerojosh
4m 58s
4m 58s
Enforce NVSHMEM minimum version of 2.6.0 to remove old workarounds. Update NVSHMEM usage guidance. Build #69: Issue comment #106 (comment) created by romerojosh
5m 39s
5m 39s
Enforce NVSHMEM minimum version of 2.6.0 to remove old workarounds. Update NVSHMEM usage guidance. Build #68: Issue comment #106 (comment) created by romerojosh
4m 25s
4m 25s
Improve support for very large inputs. Build #67: Issue comment #105 (comment) created by romerojosh
6m 10s
6m 10s
Mininum changes to compile HIP version Build #66: Issue comment #104 (comment) created by loostrum
1s
1s
Relax empty pencil check during grid descriptor creation to allow more flexibility Build #65: Issue comment #103 (comment) created by romerojosh
6m 30s
6m 30s
autotuneHaloBackend aggressively rejects valid decompositions for flat 3D grids (e.g. NxNx1) Build #64: Issue comment #101 (comment) created by romerojosh
1s
1s
Enabling support for process decompositions with empty pencils. Build #63: Issue comment #102 (comment) created by romerojosh
1s
1s
autotuneHaloBackend aggressively rejects valid decompositions for flat 3D grids (e.g. NxNx1) Build #62: Issue comment #101 (comment) created by romerojosh
1s
1s
autotuneHaloBackend aggressively rejects valid decompositions for flat 3D grids (e.g. NxNx1) Build #61: Issue comment #101 (comment) created by romerojosh
2s
2s
autotuneHaloBackend aggressively rejects valid decompositions for flat 3D grids (e.g. NxNx1) Build #60: Issue comment #101 (comment) created by romerojosh
1s
1s
0.6.1 release Build #59: Issue comment #100 (comment) created by romerojosh
8m 13s
8m 13s
Add CUDECOMP_DISABLE_MNNVL debug option. Build #58: Issue comment #99 (comment) created by romerojosh
7m 32s
7m 32s
Add CUDECOMP_DISABLE_MNNVL debug option. Build #57: Issue comment #99 (comment) created by romerojosh
8m 5s
8m 5s
Fix potential deadlock during NCCL local communicator creation. Build #56: Issue comment #98 (comment) created by romerojosh
7m 35s
7m 35s
Add SM75 support to default build with CUDA 13. Build #55: Issue comment #97 (comment) created by romerojosh
9m 3s
9m 3s
sm75_support Build #54: Issue comment #96 (comment) created by romerojosh
1s
1s
sm75_support Build #53: Issue comment #96 (comment) created by aroccon
2s
2s
sm75_support Build #52: Issue comment #96 (comment) created by romerojosh
2s
2s
0.6.0 release Build #51: Issue comment #95 (comment) created by romerojosh
7m 33s
7m 33s
Correctly propagate C++17 standard flags to NVCC compilation with older CMake versions. Build #50: Issue comment #93 (comment) created by romerojosh
9m 38s
9m 38s
Correctly propagate C++17 standard flags to NVCC compilation with older CMake versions. Build #49: Issue comment #93 (comment) created by romerojosh
6m 52s
6m 52s
Making required changes for CUDA 13 support. Build #48: Issue comment #86 (comment) created by romerojosh
10m 0s
10m 0s
Adding function for MPI communicators in Fortran integer format to simplify Fortran interfaces. Build #47: Issue comment #92 (comment) created by romerojosh
8m 49s
8m 49s
Update to Apache 2.0 license. Build #46: Issue comment #91 (comment) created by romerojosh
1s
1s