-
Notifications
You must be signed in to change notification settings - Fork 927
Pull requests: openxla/xla
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Estimate cuBLASLt fission decisions by using the dot cost model as a proxy
#48542
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
Add simplification pattern for dynamic_update_slice(pad(x), y).
#48541
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
PR #127012: Fix the Windows link by exporting the current FlatBufferFileToMlir signature
#48540
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
Refactor compiler options handling in configure.py
#48539
opened Sep 8, 2026 by
MaddipatlaChetan24
Loading…
Restrict AOT compatibility GPU testing to Nvidia H100, B200, GB200, and GB300
#48538
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
Relocate AOT compatibility GPU and CPU test targets to subpackages
#48537
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
Add tests capturing status quo of program_id, fusion IDs, and xprofilez across isomorphic modules.
#48534
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
Prevent overflow when rounding BFC allocation sizes
#48533
opened Sep 8, 2026 by
GodlyDonuts
Loading…
[ROCm ]Enable convolution_1d_test and convolution_1d_gpu_alternative_layout_test on ROCm
AMD-GPU
XLA on AMD GPU
#48531
opened Sep 8, 2026 by
linchen1-robot
Contributor
•
Draft
[XLA:GPU] Propagate oversized shared memory device property
#48528
opened Sep 8, 2026 by
dlcompilers-infra-bot
Contributor
Loading…
[XLA] Add SM 10.7 (VR200) to known CUDA compute capabilities in CI build script.
#48527
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
PR #48188: [ROCm] Use 'no_hash' string to download changing nightly/devreleases rocm tarballs
#48526
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
[Mosaic TPU] For 16-bit data, support block sizes that are smaller than
sublane*packing when large second minor is enabled. Also support static/dynamic memory transfers even when the sublane dim offset and size are not aligned to the sublane tiling.
#48525
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
Integrate LLVM at llvm/llvm-project@43574226b712
#48522
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
[ROCm] Enable dynamic_shared_memory_test on ROCm via a portable GPU kernel.
AMD-GPU
XLA on AMD GPU
#48516
opened Sep 8, 2026 by
linchen1-robot
Contributor
Loading…
[Mosaic] tpu.memref_reshape: strengthen verifier, slightly generalize and refactor
#48515
opened Sep 8, 2026 by
copybara-service
Bot
Loading…
Fix uint64 accumulation overflow in IFRT ArrayMemRegion byte-stride span
#48513
opened Sep 7, 2026 by
roopd3v
Loading…
[XLA] Refactor separate_compilation to parallel_compilation
#48512
opened Sep 7, 2026 by
copybara-service
Bot
Loading…
Integrate LLVM at llvm/llvm-project@d03786e89999
#48511
opened Sep 7, 2026 by
copybara-service
Bot
Loading…
[XLA:SDY] Do not import single-device shardings on function results in V3.
#48506
opened Sep 7, 2026 by
copybara-service
Bot
Loading…
Use canonical HloModule fingerprint for GPU profiler annotation program_id.
#48505
opened Sep 7, 2026 by
copybara-service
Bot
Loading…
Use bitmap to track input dependencies in WhileLoopSimplifier
#48504
opened Sep 7, 2026 by
copybara-service
Bot
Loading…
Previous Next
ProTip!
Adding no:label will show everything without a label.