mirrors/rocm_jax
1
0
Fork 0
mirror of https://github.com/ROCm/jax.git synced 2025-04-19 05:16:06 +00:00
Code Issues Packages Projects Releases Wiki Activity
rocm_jax/jaxlib/gpu
History
George Necula 7d452adfd3 Add support for dynamic shapes to GPU threefry2x32 custom call. 
In presence of dynamic shapes the ThreeFry2x32Descriptor will contain the
value n=-1, and the actual desired output length will be passed as
an additional operand. If the shape is static then the length will be
passed as part of the descriptor.

PiperOrigin-RevId: 497945778
2022-12-27 04:48:26 -08:00
..
blas_kernels.cc
Merge CUDA and ROCM kernel code in jaxlib.
2022-10-25 07:23:34 -07:00
blas_kernels.h
Merge CUDA and ROCM kernel code in jaxlib.
2022-10-25 07:23:34 -07:00
blas.cc
Remove an unnecessary include/ from pybind11 include paths.
2022-11-30 14:20:02 -08:00
BUILD
Add CUDNN custom call for LSTM. Exposed as jax.experimental.rnn module.
2022-11-28 14:31:48 -08:00
gpu_kernel_helpers.cc
Rollback of "Add CUDNN custom call for LSTM. Exposed as jax.experimental.rnn module."
2022-11-23 07:48:05 -08:00
gpu_kernel_helpers.h
Rollback of "Add CUDNN custom call for LSTM. Exposed as jax.experimental.rnn module."
2022-11-23 07:48:05 -08:00
gpu_kernels.cc
Remove unused potrf kernels.
2022-11-17 15:22:13 -08:00
linalg.cc
Remove an unnecessary include/ from pybind11 include paths.
2022-11-30 14:20:02 -08:00
lu_pivot_kernels.cc
Merge CUDA and ROCM kernel code in jaxlib.
2022-10-25 07:23:34 -07:00
lu_pivot_kernels.cu.cc
Merge CUDA and ROCM kernel code in jaxlib.
2022-10-25 07:23:34 -07:00
lu_pivot_kernels.h
Merge CUDA and ROCM kernel code in jaxlib.
2022-10-25 07:23:34 -07:00
prng_kernels.cc
Merge CUDA and ROCM kernel code in jaxlib.
2022-10-25 07:23:34 -07:00
prng_kernels.cu.cc
Add support for dynamic shapes to GPU threefry2x32 custom call.
2022-12-27 04:48:26 -08:00
prng_kernels.h
Add support for dynamic shapes to GPU threefry2x32 custom call.
2022-12-27 04:48:26 -08:00
prng.cc
Remove an unnecessary include/ from pybind11 include paths.
2022-11-30 14:20:02 -08:00
rnn_kernels.cc
Add CUDNN custom call for LSTM. Exposed as jax.experimental.rnn module.
2022-11-28 14:31:48 -08:00
rnn_kernels.h
Add CUDNN custom call for LSTM. Exposed as jax.experimental.rnn module.
2022-11-28 14:31:48 -08:00
rnn.cc
Remove an unnecessary include/ from pybind11 include paths.
2022-11-30 14:20:02 -08:00
solver_kernels.cc
Remove unused potrf kernels.
2022-11-17 15:22:13 -08:00
solver_kernels.h
Remove unused potrf kernels.
2022-11-17 15:22:13 -08:00
solver.cc
Remove an unnecessary include/ from pybind11 include paths.
2022-11-30 14:20:02 -08:00
sparse_kernels.cc
[sparse] Update the default CUSPARSE SPMV and SPMM algorithms in jaxlib.
2022-11-01 16:01:50 -07:00
sparse_kernels.h
Merge CUDA and ROCM kernel code in jaxlib.
2022-10-25 07:23:34 -07:00
sparse.cc
Remove an unnecessary include/ from pybind11 include paths.
2022-11-30 14:20:02 -08:00
vendor.h
Add CUDNN custom call for LSTM. Exposed as jax.experimental.rnn module.
2022-11-28 14:31:48 -08:00