-
Notifications
You must be signed in to change notification settings - Fork 75k
Pull requests: tensorflow/tensorflow
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[XLA:GPU] Exclude Intel targets from internal wildcard build
#103593
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
[XLA:GPU][XTile] Support 0D tensors in emitting transpose, bitcast, broadcast and dot helpers.
#103592
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
[XLA:GPU/TMA] Move the restriction on GEMMs with broadcasts and pipelining from the autotuner to the emitter.
#103589
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
Replace forward quotes with backward quotes in TFLite opaque api documentation.
#103588
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
[XLA:GPU]: Add a kTritonCollectiveFusion kind.
#103587
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
[XLA:GPU] Add collective emitter to triton/ codegen.
#103586
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
Add flag that dumps a serialized
GpuExecutable
#103585
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
Improve kernel loader debugging in cuda executor
#103584
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
PR #33504: Add human-readable units to "Out of memory" errors (e.g. 7.34GB)
#103581
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
[XLA:GPU] Pass peer pointers for each kernel parameter.
#103580
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
Rename PjRtTopologyDescription methods from "PerProcess" to "PerHost".
#103577
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
Rename
PjRtTopologyDescription::ProcessCount() to HostCount().
#103576
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
Preserve tmpdir for rocm. Fixes build for certain docker setups that don't have standard tmpdir locations
#103575
opened Nov 5, 2025 by
copybara-service
bot
•
Draft
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.