mirror of
https://github.com/ROCm/jax.git
synced 2025-04-17 04:16:07 +00:00
5e915d3307
Unlike the other more detailed ports, this version doesn't take full advantage of the features provided by the FFI. For example, it would be possible to update the kernels to use the ScratchAllocator instead of querying the workspace size during lowering. However, since these kernels are really only meant to be experimental, it's not obvious to me that it's worth the extra work to do anything more sophisticated. PiperOrigin-RevId: 724016331
93 lines
2.1 KiB
Python
93 lines
2.1 KiB
Python
# Copyright 2018 The JAX Authors.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# https://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
# Shared CUDA/ROCM GPU kernels.
|
|
|
|
load(
|
|
"//jaxlib:jax.bzl",
|
|
"cc_proto_library",
|
|
"jax_visibility",
|
|
"xla_py_proto_library",
|
|
)
|
|
# Placeholder: load proto_library
|
|
|
|
licenses(["notice"])
|
|
|
|
package(
|
|
default_applicable_licenses = [],
|
|
default_visibility = ["//jax:internal"],
|
|
)
|
|
|
|
exports_files(srcs = [
|
|
"blas.cc",
|
|
"blas_handle_pool.cc",
|
|
"blas_handle_pool.h",
|
|
"blas_kernels.cc",
|
|
"blas_kernels.h",
|
|
"ffi_wrapper.h",
|
|
"gpu_kernel_helpers.cc",
|
|
"gpu_kernel_helpers.h",
|
|
"gpu_kernels.cc",
|
|
"hybrid.cc",
|
|
"hybrid_kernels.cc",
|
|
"hybrid_kernels.h",
|
|
"linalg.cc",
|
|
"linalg_kernels.cc",
|
|
"linalg_kernels.cu.cc",
|
|
"linalg_kernels.h",
|
|
"make_batch_pointers.cu.cc",
|
|
"make_batch_pointers.h",
|
|
"prng.cc",
|
|
"prng_kernels.cc",
|
|
"prng_kernels.cu.cc",
|
|
"prng_kernels.h",
|
|
"rnn.cc",
|
|
"rnn_kernels.cc",
|
|
"rnn_kernels.h",
|
|
"solver.cc",
|
|
"solver_handle_pool.cc",
|
|
"solver_handle_pool.h",
|
|
"solver_interface.cc",
|
|
"solver_interface.h",
|
|
"solver_kernels.cc",
|
|
"solver_kernels.h",
|
|
"solver_kernels_ffi.cc",
|
|
"solver_kernels_ffi.h",
|
|
"sparse.cc",
|
|
"sparse_kernels.cc",
|
|
"sparse_kernels.h",
|
|
"triton.cc",
|
|
"triton_kernels.cc",
|
|
"triton_kernels.h",
|
|
"triton_utils.cc",
|
|
"triton_utils.h",
|
|
"vendor.h",
|
|
])
|
|
|
|
proto_library(
|
|
name = "triton_proto",
|
|
srcs = ["triton.proto"],
|
|
)
|
|
|
|
cc_proto_library(
|
|
name = "triton_cc_proto",
|
|
deps = [":triton_proto"],
|
|
)
|
|
|
|
xla_py_proto_library(
|
|
name = "triton_py_pb2",
|
|
visibility = jax_visibility("triton_proto_py_users"),
|
|
deps = [":triton_proto"],
|
|
)
|