Skip to content

Pull requests: openxla/xla

New pull request New
676 Open 44,387 Closed
676 Open 44,387 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[XLA:GPU]: Propagate XLA flags on compilation reload.
#45802 opened Jul 16, 2026 by copybara-service Bot Loading…
@sohaibiftikhar
PR #45595: [XLA:GPU] Support VA remapping of parameter buffer allocations
#45801 opened Jul 16, 2026 by copybara-service Bot Loading…
Add unreduced rule for reduce_max and reduce_min
#45800 opened Jul 16, 2026 by copybara-service Bot Loading…
[GPU] Reroute external hero users to the fusion output in DSFv2 rewriter
#45799 opened Jul 16, 2026 by terryysun Contributor Loading…
10
[XLA:GPU] When Triton GEMM is disabled, normalize dots to have at most one non-contracting dimensions from either side so that cuBLAS could take it.
#45798 opened Jul 16, 2026 by copybara-service Bot Loading…
@mooskagh
[XLA:GPU][codegen] Add SPIR-V barrier type for codegen tests
#45797 opened Jul 16, 2026 by akhilgoe Contributor Loading…
Automated Code Change
#45796 opened Jul 16, 2026 by copybara-service Bot Loading…
Keep foreign input buffers alive until execution completion under the compute-synchronized allocation model
#45795 opened Jul 16, 2026 by copybara-service Bot Loading…
Update replica group comparator
#45794 opened Jul 16, 2026 by copybara-service Bot Loading…
@felixwqp
Reverts 1c764a9d6d803887aadb7f566ed7b51c453bd07b
#45793 opened Jul 16, 2026 by copybara-service Bot Loading…
@pineapplejuice233
Add multi-chunk vector overloads for ExportToTensorBoard and SaveXSpaceChunks to support >2GB TPU profiles in JAX and robust XSpace record persistence.
#45792 opened Jul 16, 2026 by copybara-service Bot Loading…
@adityakakarot
Automated Code Change
#45788 opened Jul 16, 2026 by copybara-service Bot Loading…
Automated Code Change
#45787 opened Jul 16, 2026 by copybara-service Bot Loading…
Automated Code Change
#45786 opened Jul 16, 2026 by copybara-service Bot Loading…
Rename IsDefaultLayout to IsDefaultPriority.
#45782 opened Jul 15, 2026 by copybara-service Bot Loading…
Bump rules_ml_toolchain to version with C++ modules support
#45781 opened Jul 15, 2026 by copybara-service Bot Loading…
1
[PJRT] Roll forward UndonatableCommonPjrtBuffer with fix.
#45780 opened Jul 15, 2026 by copybara-service Bot Loading…
Update TPU SDK test export configuration for accel_ssw batch tests.
#45779 opened Jul 15, 2026 by copybara-service Bot Loading…
[XLA:MSA] Allow scalars to be pinned to alternate memory in MSA.
#45778 opened Jul 15, 2026 by copybara-service Bot Loading…
@subhankarshah
Support saving greater than 2GB TPU profiles in JAX in Riegeli format
#45777 opened Jul 15, 2026 by copybara-service Bot Loading…
@adityakakarot
Reverts a51c881bf6e9aa01f6db0d3b3b29e973b0eec109
#45776 opened Jul 15, 2026 by copybara-service Bot Loading…
@beckerhe
[XLA] Externalize XLA HLO dump utilities to third_party/tensorflow.
#45775 opened Jul 15, 2026 by copybara-service Bot Loading…
[XLA:GPU] Rewrite seeded and live carry add scans to CUB device scans.
#45774 opened Jul 15, 2026 by copybara-service Bot Loading…
@patrick-toulme
[WIP] Expose NCCL Collective Resources from FFI
#45772 opened Jul 15, 2026 by mgoldfarb-nvidia Contributor • Draft
3
Properly move iota_ (std::move(other.iota_)) and null out other.array_ (std::exchange(other.array_, nullptr)) inside TileAssignment move constructor and move assignment operator.
#45770 opened Jul 15, 2026 by copybara-service Bot Loading…
@wsmoses
ProTip! Updated in the last three days: updated:>2026-07-13.