-
Notifications
You must be signed in to change notification settings - Fork 782
Pull requests: openxla/xla
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
PR #39201: [xla:gpu] Split legacy custom calls into LegacyCustomCallThunk
#41072
opened Apr 17, 2026 by
copybara-service
bot
Loading…
PR #40967: [xla:gpu] Cache NCCL Communicator properties
#41071
opened Apr 17, 2026 by
copybara-service
bot
Loading…
[xla:gpu] Unify CUDA allocators under cuMemCreate allocator
#41069
opened Apr 16, 2026 by
ezhulenev
Contributor
Loading…
[XLA:TPU] BitcastConvert updated for packed
preds
#41068
opened Apr 16, 2026 by
copybara-service
bot
Loading…
convert FusionInterface to return AsyncThunkSequence
#41065
opened Apr 16, 2026 by
copybara-service
bot
Loading…
- Add UpdateReachabilityThroughInstructions to HloReachabilityMap to support batched updates.
#41064
opened Apr 16, 2026 by
copybara-service
bot
Loading…
Add subprocess registration hooks to JAX profiler libraries.
#41063
opened Apr 16, 2026 by
copybara-service
bot
Loading…
Add support for variadic arguments in AsyncUpdate and related tests.
#41061
opened Apr 16, 2026 by
copybara-service
bot
Loading…
[XLA:GPU] fix the reserve call in experimental tile
#41060
opened Apr 16, 2026 by
copybara-service
bot
Loading…
This CL introduce a new tie breaker for two
async-done:
#41059
opened Apr 16, 2026 by
copybara-service
bot
Loading…
PR #40917: [xla:gpu] Print HLO/Executable fingerprints when loading from AOT result
#41057
opened Apr 16, 2026 by
copybara-service
bot
Loading…
Cross Host Data Transfers: Add CrossHostTransferBuffers; modify CrossHostSendBuffers to call into it
#41055
opened Apr 16, 2026 by
rao-ashish
Contributor
Loading…
[XLA:GPU] Remove cutlass & custom kernel fusions from XLA:GPU.
#41054
opened Apr 16, 2026 by
copybara-service
bot
Loading…
[XLA:GPU][ONEAPI] Oneapi plugin support for Intel GPU (part1) - sycl name changes to oneapi
#41051
opened Apr 16, 2026 by
ashiqimranintel
Contributor
Loading…
Rollback PR #37936: since it breaks ragged_collective_test_gpu in JAX when machine have >2 GPUs.
#41049
opened Apr 16, 2026 by
copybara-service
bot
Loading…
PR #40920: [xla] Add a separate pass to propagate metadata across kCall instruction
#41048
opened Apr 16, 2026 by
copybara-service
bot
Loading…
Use debug options from flags instead of
ExecutableBuildOptions to extract HloShardingV3 flag because we are planning to enable/disable this flag by default in debug options and if some JAX users wants to disable they can do it via XLA_FLAGS
#41047
opened Apr 16, 2026 by
copybara-service
bot
Loading…
[XLA:GPU] Remove the FusionDecision ugliness.
#41046
opened Apr 16, 2026 by
copybara-service
bot
Loading…
Fix alignment logic in CUB scratch size logic
#41043
opened Apr 16, 2026 by
copybara-service
bot
Loading…
[XLA:GPU]: Add some debugging information for triton collective kernels
#41041
opened Apr 16, 2026 by
copybara-service
bot
Loading…
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.