jax
CI: 11/14/24 upstream sync
#142
Merged

Commits
  • Re-enable cudnn_fusion_test on A100.
    sergachev committed 1 year ago
  • Internal: create decorators for defining ufuncs
    jakevdp committed 1 year ago
  • Allow setting default_device with platform names.
    Stella-S-Yan committed 1 year ago
  • jax.scipy.linalg.toeplitz: support implicit batching
    jakevdp committed 1 year ago
  • Update some outdated syntax in FFI tutorial.
    dfm committed 1 year ago
  • Add an example FFI call to demonstrate the use of global state.
    dfm committed 1 year ago
  • Add an option to specify mock GPU topology
    nvjax committed 1 year ago
  • Add GPU overflow bugfix (#24846) to changelog.
    dfm committed 1 year ago
  • Merge pull request #24860 from dfm:overflow-changelog
    Google-ML-Automation committed 1 year ago
  • Fix pgle profiling, broken in previous change.
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24806 from jakevdp:ufunc-decorator
    Google-ML-Automation committed 1 year ago
  • Specialize ufunc.reduce for monoidal binary ufuncs.
    jakevdp committed 1 year ago
  • Change to internal dead code elimination. Now the functions in `dce_rules` are responsible for checking if the equation has no used outputs or effects, and behaving appropriately in that case (which usually means eliminating said equation).
    Google-ML-Automation committed 1 year ago
  • [Mosaic] Fix mask creation for packed sublanes
    Google-ML-Automation committed 1 year ago
  • [mosaic_gpu] The profiler now uses FFI calls for creating events and computing elapsed time
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24792 from dfm:ffi-tutorial-outdated
    Google-ML-Automation committed 1 year ago
  • Update XLA dependency to use revision
    Google-ML-Automation committed 1 year ago
  • Change the assumed width of the bool packing in the early-lowering checks in pallas
    Google-ML-Automation committed 1 year ago
  • [Mosaic] Add parameter names to tpu.sem_signal and add tests
    Google-ML-Automation committed 1 year ago
  • Dedent your yields!
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24842 from jakevdp:batched-toeplitz
    Google-ML-Automation committed 1 year ago
  • Add new CI scripts for running Bazel CPU presubmits
    Google-ML-Automation committed 1 year ago
  • [XLA] Make our LLVM usage more googley
    Google-ML-Automation committed 1 year ago
  • Fix parallel pgle-tests execution.
    Google-ML-Automation committed 1 year ago
  • Add a shard map replication rule for cond_p.
    dfm committed 1 year ago
  • Avoid using a contextmanager in Primitive.bind.
    Google-ML-Automation committed 1 year ago
  • Fix bug where the Python wrapper to ParseArguments() didn't intern the static argnames strings, causing false mismatches when searching for static arguments.
    Google-ML-Automation committed 1 year ago
  • [AutoPGLE] Fix test after pjrt cache refactoring
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24877 from dfm:cond-rep-rule
    Google-ML-Automation committed 1 year ago
  • Add square_p
    pearu committed 1 year ago
  • Add Pallas Triton lowering for jax.lax.bitcast_convert_type.
    Google-ML-Automation committed 1 year ago
  • When caching is enabled, also enable XLA caching features as well
    trevor-m committed 1 year ago
  • Merge pull request #24069 from sergachev:cudnn_fusion_test_a100
    Google-ML-Automation committed 1 year ago
  • Merge pull request #22899 from trevor-m:cache
    Google-ML-Automation committed 1 year ago
  • doc: link directly to installation on the main page
    jakevdp committed 1 year ago
  • Fix typos: Change 'arugments' to 'arguments'.
    carlosgmartin committed 1 year ago
  • Merge pull request #24874 from pearu:pearu/square_p
    Google-ML-Automation committed 1 year ago
  • Update XLA dependency to use revision
    Google-ML-Automation committed 1 year ago
  • Stop using generators for linear_util transformations.
    dougalm committed 1 year ago
  • Don't override --xla_tpu_use_enhanced_launch_barrier if explicitly set
    skye committed 1 year ago
  • Merge pull request #24883 from jakevdp:doc-installation
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24886 from carlosgmartin:fix_typos
    Google-ML-Automation committed 1 year ago
  • Avoid repeatedly rebuilding a tuple in issubdtype.
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24810 from jakevdp:bitwise-reduce
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24873 from jax-ml:no-gen-linear-util
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24889 from skye:barrier
    Google-ML-Automation committed 1 year ago
  • Merge pull request #23805 from dfm:ffi-examples-state
    Google-ML-Automation committed 1 year ago
  • [pallas:mosaic_gpu] Ported two pipelining optimizations to `emit_pipeline`
    Google-ML-Automation committed 1 year ago
  • Make the jaxpr for jnp.pad in "constant" mode more succinct.
    hawkinsp committed 1 year ago
  • Don't use an out-of-line lowering for integer_pow for small powers.
    hawkinsp committed 1 year ago
  • Merge pull request #23374 from jaro-sevcik:mock-topology-config
    Google-ML-Automation committed 1 year ago
  • Fixed typo in asan.yaml
    vfdev-5 committed 1 year ago
  • finalize deprecation of jax.clear_backends
    jakevdp committed 1 year ago
  • Merge pull request #24899 from vfdev-5:patch-2
    Google-ML-Automation committed 1 year ago
  • Fix pgle test handling of directory removal.
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24865 from jakevdp:dep-clear-backends
    Google-ML-Automation committed 1 year ago
  • [Mosaic][TPU] Omit short circuiting of relayout (we should always relayout!) and implement product mismatch case for where we relayout from replicated to offset, and the number of vregs changes.
    Google-ML-Automation committed 1 year ago
  • [sharding_in_types] Support shard_map with sharding in types. Right now only full manual mode is supported.
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24751 from Stella-S-Yan:feature/default_device_str
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24896 from hawkinsp:pad
    Google-ML-Automation committed 1 year ago
  • Merge pull request #24897 from hawkinsp:ipow
    Google-ML-Automation committed 1 year ago
  • Merge remote-tracking branch 'origin/rocm-main' into ci-upstream-sync-25_1
    invalid-email-address committed 1 year ago
  • Add commit to see if it triggers CI
    charleshofer committed 1 year ago
Loading