openxla / xla

A machine learning compiler for GPUs, CPUs, and ML accelerators

openxla/xla Issues

build error
Closed 6 days ago5
[XLA:GPU] cuDNN norm rewriter fails with OOB shape access
Updated 6 hours ago5
Question : When do we not honor "preferred_element_type" ?
Updated 6 hours ago
Wrong output from JAX test after onednn change
Closed a day ago4
ReduceScatterCreator - Misses One Pattern
Updated a day ago2
SPMD - Handle Gather/Scatter is Harcoded to IndexParallel
Updated a day ago1
All Gather Combiner - Coalesces different dtypes
Updated a day ago1
lowering stablehlo to ptx
Closed 7 days ago5
Support builds with cuDNN 9
Updated 2 days ago1
How are random seeds set for RNG operations?
Updated 2 days ago3
HandleReshape in SPMD partitioner splits sharding in some cases where it is unoptimal
Updated 2 days ago1
Interleaved/Looping SPMD Pipelines
Closed 5 days ago1
Build from source fails
Updated 6 days ago7
The official website documentation and the repository documentation do not correspond
Updated 6 days ago5
PJRT `CopyCpuBufferToLiteral` of JAX buffer taking too long
Closed 21 days ago9
Compiling scatter results in very slow while-loop on TPU
Updated 13 days ago
Compiling xla/mlir/tools/mlir_interpreter/dialects/util.cc failed
Closed 14 days ago3
MHLO Extraction from XLA Compiler
Updated 15 days ago4
Controlling a single compiler pass in XLA for CPU target
Updated 15 days ago2
Porting XLA to different backends.
Updated 19 days ago4
BF16 matmul slower than F32 matmul on T4 GPU
Updated 21 days ago3
Build fails with ROCm on Gentoo Linux
Updated 21 days ago12
OpenCL Support.
Updated 22 days ago2
[xla:auto_sharding] Question about resharding costs of Reshape strategies
Updated 23 days ago
Workable example on normal DNN model
Updated a month ago
[Feature Request]Add more comm op support on gpu_hlo_cost_analysis
Updated a month ago4
gpu_fused_mha_test fails at HEAD on H100
Closed a month ago1
How to build for TPU ?
Updated a month ago8
"Could not find executable `nvidia-smi`" for `./configure.py --backend=CUDA`
Updated a month ago6
[xla:GPU] Is gpu_collective_performance_model correct?
Closed a month ago9
`PjRtCApiBuffer::AcquireExternalReference` does not support `WaitUntilBufferReadyOnStream`
Closed a month ago2
XLA documentation for Windows
Updated a month ago1
Implement GitHub Presubmit Checks for Windows Environment
Updated a month ago1
gpu f16 cast to fp32 calculation, and then converted back?
Updated a month ago1
Integer overflow in ZlibOutputBuffer causes crash
Updated a month ago3
gpu_hlo_cost_analysis NumOfDevices always return 0
Closed a month ago2
XLA CollectiveOpsUtil does not include kReduceScatter
Closed a month ago2
call ptxas become defunct, cause xla hung
Updated a month ago2
[XLA:CPU] oneDNN Softmax gives inaccurate results
Updated a month ago1
Compiling with `build --config nonccl` fails
Updated 2 months ago1
No static version for linking NCCL (libnccl)
Updated 2 months ago2
cuDNN fMHA failures return unhelpful error messages
Updated 2 months ago
Incorrect boundary behaviour of `igamma`, `igammac`
Closed 2 months ago10
GpuLayoutAssignment pass unexpectedly replaced the host memory space color of the copy-start/copy-done instructions
Closed 2 months ago6
Is there any operation in XLA capable of reserving blank memory?
Closed 2 months ago2
Help linking transitive dependencies when building shared library with XLA
Closed 2 months ago1
Redefined macros due to use of Abseil and TSL logging
Updated 2 months ago
CollectiveBroadcast should add `use_global_device_ids` to its hlo construction
Updated 2 months ago
Potentially Misbehaving GPU
Updated 3 months ago8
XLA Backend support for third party AI Accelerator
Updated 3 months ago