Skip to content

Actions: openxla/shardy

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
2,213 workflow runs
2,213 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

Refactor to separate finding axes without considering replicated factors first. Buildifier #1144: Pull request #390 synchronize by copybara-service bot
February 19, 2025 13:45 20s test_728598822
test_728598822
February 19, 2025 13:45 20s
Refactor to separate finding axes without considering replicated factors first. Build/Test Bazel #977: Pull request #390 synchronize by copybara-service bot
February 19, 2025 13:45 7m 45s test_728598822
test_728598822
February 19, 2025 13:45 7m 45s
Partially handle operations with permutation factors for explicit reshards. Build/Test Bazel #976: Pull request #387 synchronize by copybara-service bot
February 19, 2025 12:58 7m 42s test_727841913
test_727841913
February 19, 2025 12:58 7m 42s
Partially handle operations with permutation factors for explicit reshards. Buildifier #1143: Pull request #387 synchronize by copybara-service bot
February 19, 2025 12:58 25s test_727841913
test_727841913
February 19, 2025 12:58 25s
Refactor to separate finding axes without considering replicated factors first. Build/Test Bazel #975: Pull request #390 opened by copybara-service bot
February 19, 2025 11:46 7m 38s test_728598822
test_728598822
February 19, 2025 11:46 7m 38s
Refactor to separate finding axes without considering replicated factors first. Buildifier #1142: Pull request #390 opened by copybara-service bot
February 19, 2025 11:46 22s test_728598822
test_728598822
February 19, 2025 11:46 22s
#sdy For a future all-to-all from source dimension d1 to target dimension d2, slice available axes on d1 instead d2 to avoid a redundant collective-permute if certain conditions are met. Buildifier #1141: Pull request #389 synchronize by copybara-service bot
February 19, 2025 10:27 24s test_723874369
test_723874369
February 19, 2025 10:27 24s
#sdy For a future all-to-all from source dimension d1 to target dimension d2, slice available axes on d1 instead d2 to avoid a redundant collective-permute if certain conditions are met. Build/Test Bazel #974: Pull request #389 synchronize by copybara-service bot
February 19, 2025 10:27 7m 37s test_723874369
test_723874369
February 19, 2025 10:27 7m 37s
#sdy For a future all-to-all from source dimension d1 to target dimension d2, slice available axes on d1 instead d2 to avoid a redundant collective-permute if certain conditions are met. Buildifier #1140: Pull request #389 synchronize by copybara-service bot
February 18, 2025 17:31 36s test_723874369
test_723874369
February 18, 2025 17:31 36s
#sdy For a future all-to-all from source dimension d1 to target dimension d2, slice available axes on d1 instead d2 to avoid a redundant collective-permute if certain conditions are met. Build/Test Bazel #973: Pull request #389 synchronize by copybara-service bot
February 18, 2025 17:31 7m 43s test_723874369
test_723874369
February 18, 2025 17:31 7m 43s
#sdy For a future all-to-all from source dimension d1 to target dimension d2, slice available axes on d1 instead d2 to avoid a redundant collective-permute if certain conditions are met. Buildifier #1139: Pull request #389 opened by copybara-service bot
February 18, 2025 17:15 21s test_723874369
test_723874369
February 18, 2025 17:15 21s
#sdy For a future all-to-all from source dimension d1 to target dimension d2, slice available axes on d1 instead d2 to avoid a redundant collective-permute if certain conditions are met. Build/Test Bazel #972: Pull request #389 opened by copybara-service bot
February 18, 2025 17:15 7m 48s test_723874369
test_723874369
February 18, 2025 17:15 7m 48s
#sdy Add an event to the op priority schedule to propagate through all ops after element-wise ops, but **only** through non-passthrough factors. Buildifier #1138: Pull request #384 synchronize by copybara-service bot
February 18, 2025 16:50 20s test_726942878
test_726942878
February 18, 2025 16:50 20s
#sdy Add an event to the op priority schedule to propagate through all ops after element-wise ops, but **only** through non-passthrough factors. Build/Test Bazel #971: Pull request #384 synchronize by copybara-service bot
February 18, 2025 16:50 7m 27s test_726942878
test_726942878
February 18, 2025 16:50 7m 27s
During sharding group op during import, emit a warning instead of error if the group has a conflicting initial sharding. However, if there is a conflict: Build/Test Bazel #970: Pull request #388 opened by copybara-service bot
February 18, 2025 13:58 7m 6s test_727888959
test_727888959
February 18, 2025 13:58 7m 6s
During sharding group op during import, emit a warning instead of error if the group has a conflicting initial sharding. However, if there is a conflict: Buildifier #1137: Pull request #388 opened by copybara-service bot
February 18, 2025 13:58 30s test_727888959
test_727888959
February 18, 2025 13:58 30s
#sdy when inserting all-to-alls, first try to insert all-to-alls that move axes to the right position at the target dimension. Buildifier #1136: Pull request #386 synchronize by copybara-service bot
February 17, 2025 16:08 38s test_727822196
test_727822196
February 17, 2025 16:08 38s
#sdy when inserting all-to-alls, first try to insert all-to-alls that move axes to the right position at the target dimension. Build/Test Bazel #969: Pull request #386 synchronize by copybara-service bot
February 17, 2025 16:08 7m 21s test_727822196
test_727822196
February 17, 2025 16:08 7m 21s
Partially handle operations with permutation factors for explicit reshards. Build/Test Bazel #968: Pull request #387 opened by copybara-service bot
February 17, 2025 14:32 7m 19s test_727841913
test_727841913
February 17, 2025 14:32 7m 19s
Partially handle operations with permutation factors for explicit reshards. Buildifier #1135: Pull request #387 opened by copybara-service bot
February 17, 2025 14:32 31s test_727841913
test_727841913
February 17, 2025 14:32 31s
#sdy when inserting all-to-alls, first try to insert all-to-alls that move axes to the right position at the target dimension. Buildifier #1134: Pull request #386 synchronize by copybara-service bot
February 17, 2025 13:23 24s test_727822196
test_727822196
February 17, 2025 13:23 24s
#sdy when inserting all-to-alls, first try to insert all-to-alls that move axes to the right position at the target dimension. Build/Test Bazel #967: Pull request #386 synchronize by copybara-service bot
February 17, 2025 13:23 7m 10s test_727822196
test_727822196
February 17, 2025 13:23 7m 10s
#sdy when inserting all-to-alls, first try to insert all-to-alls that move axes to the right position at the target dimension. Buildifier #1133: Pull request #386 opened by copybara-service bot
February 17, 2025 13:10 21s test_727822196
test_727822196
February 17, 2025 13:10 21s
#sdy when inserting all-to-alls, first try to insert all-to-alls that move axes to the right position at the target dimension. Build/Test Bazel #966: Pull request #386 opened by copybara-service bot
February 17, 2025 13:10 7m 10s test_727822196
test_727822196
February 17, 2025 13:10 7m 10s
Tag the reverse dimensions as kPermutation, since the SPMD partitioner adds halo exchange if the reverse dimensions is sharded. Buildifier #1132: Pull request #385 synchronize by copybara-service bot
February 14, 2025 22:14 25s test_727044981
test_727044981
February 14, 2025 22:14 25s