mirror of
https://github.com/ROCm/jax.git
synced 2025-04-16 11:56:07 +00:00
378 lines
10 KiB
Python
378 lines
10 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.
|
|
|
|
# NVIDIA CUDA kernels
|
|
|
|
load(
|
|
"//jaxlib:jax.bzl",
|
|
"cuda_library",
|
|
"if_cuda_is_configured",
|
|
"pybind_extension",
|
|
)
|
|
|
|
licenses(["notice"])
|
|
|
|
package(default_visibility = ["//:__subpackages__"])
|
|
|
|
cc_library(
|
|
name = "cuda_vendor",
|
|
hdrs = [
|
|
"//jaxlib/gpu:vendor.h",
|
|
],
|
|
defines = ["JAX_GPU_CUDA=1"],
|
|
deps = [
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
"@local_config_cuda//cuda:cudnn_header",
|
|
],
|
|
)
|
|
|
|
cc_library(
|
|
name = "cuda_gpu_kernel_helpers",
|
|
srcs = [
|
|
"//jaxlib/gpu:gpu_kernel_helpers.cc",
|
|
],
|
|
hdrs = [
|
|
"//jaxlib/gpu:gpu_kernel_helpers.h",
|
|
],
|
|
copts = [
|
|
"-fexceptions",
|
|
],
|
|
features = ["-use_header_modules"],
|
|
deps = [
|
|
":cuda_vendor",
|
|
"@xla///stream_executor/cuda:cusolver_lib",
|
|
"@xla///stream_executor/cuda:cusparse_lib",
|
|
"@com_google_absl//absl/memory",
|
|
"@com_google_absl//absl/status",
|
|
"@com_google_absl//absl/status:statusor",
|
|
"@com_google_absl//absl/strings",
|
|
"@com_google_absl//absl/strings:str_format",
|
|
"@local_config_cuda//cuda:cublas_headers",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
],
|
|
)
|
|
|
|
cc_library(
|
|
name = "cublas_kernels",
|
|
srcs = ["//jaxlib/gpu:blas_kernels.cc"],
|
|
hdrs = ["//jaxlib/gpu:blas_kernels.h"],
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_vendor",
|
|
"//jaxlib:handle_pool",
|
|
"//jaxlib:kernel_helpers",
|
|
"@xla///service:custom_call_status",
|
|
"@xla///stream_executor/cuda:cublas_lib",
|
|
"@xla///stream_executor/cuda:cudart_stub",
|
|
"@com_google_absl//absl/algorithm:container",
|
|
"@com_google_absl//absl/base",
|
|
"@com_google_absl//absl/base:core_headers",
|
|
"@com_google_absl//absl/hash",
|
|
"@com_google_absl//absl/memory",
|
|
"@com_google_absl//absl/status",
|
|
"@com_google_absl//absl/strings",
|
|
"@com_google_absl//absl/strings:str_format",
|
|
"@com_google_absl//absl/synchronization",
|
|
"@local_config_cuda//cuda:cublas_headers",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
],
|
|
)
|
|
|
|
pybind_extension(
|
|
name = "_blas",
|
|
srcs = ["//jaxlib/gpu:blas.cc"],
|
|
copts = [
|
|
"-fexceptions",
|
|
"-fno-strict-aliasing",
|
|
],
|
|
features = ["-use_header_modules"],
|
|
module_name = "_blas",
|
|
deps = [
|
|
":cublas_kernels",
|
|
":cuda_vendor",
|
|
"//jaxlib:kernel_pybind11_helpers",
|
|
"@xla///stream_executor/cuda:cublas_lib",
|
|
"@com_google_absl//absl/container:flat_hash_map",
|
|
"@com_google_absl//absl/strings:str_format",
|
|
"@pybind11",
|
|
],
|
|
)
|
|
|
|
cc_library(
|
|
name = "cudnn_rnn_kernels",
|
|
srcs = ["//jaxlib/gpu:rnn_kernels.cc"],
|
|
hdrs = ["//jaxlib/gpu:rnn_kernels.h"],
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_vendor",
|
|
"//jaxlib:handle_pool",
|
|
"//jaxlib:kernel_helpers",
|
|
"@xla///service:custom_call_status",
|
|
"@xla///stream_executor/cuda:cudart_stub",
|
|
"@xla///stream_executor/cuda:cudnn_lib",
|
|
"@com_google_absl//absl/status",
|
|
"@com_google_absl//absl/status:statusor",
|
|
"@com_google_absl//absl/strings:str_format",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
],
|
|
)
|
|
|
|
pybind_extension(
|
|
name = "_rnn",
|
|
srcs = ["//jaxlib/gpu:rnn.cc"],
|
|
copts = [
|
|
"-fexceptions",
|
|
"-fno-strict-aliasing",
|
|
],
|
|
features = ["-use_header_modules"],
|
|
module_name = "_rnn",
|
|
deps = [
|
|
":cuda_vendor",
|
|
":cudnn_rnn_kernels",
|
|
"//jaxlib:kernel_pybind11_helpers",
|
|
"@com_google_absl//absl/container:flat_hash_map",
|
|
"@com_google_absl//absl/strings:str_format",
|
|
"@pybind11",
|
|
"@pybind11_abseil//pybind11_abseil:status_casters",
|
|
],
|
|
)
|
|
|
|
cc_library(
|
|
name = "cusolver_kernels",
|
|
srcs = ["//jaxlib/gpu:solver_kernels.cc"],
|
|
hdrs = ["//jaxlib/gpu:solver_kernels.h"],
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_vendor",
|
|
"//jaxlib:handle_pool",
|
|
"//jaxlib:kernel_helpers",
|
|
"@xla///service:custom_call_status",
|
|
"@xla///stream_executor/cuda:cusolver_lib",
|
|
"@com_google_absl//absl/status",
|
|
"@com_google_absl//absl/status:statusor",
|
|
"@com_google_absl//absl/synchronization",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
],
|
|
)
|
|
|
|
pybind_extension(
|
|
name = "_solver",
|
|
srcs = ["//jaxlib/gpu:solver.cc"],
|
|
copts = [
|
|
"-fexceptions",
|
|
"-fno-strict-aliasing",
|
|
],
|
|
features = ["-use_header_modules"],
|
|
module_name = "_solver",
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_vendor",
|
|
":cusolver_kernels",
|
|
"//jaxlib:kernel_pybind11_helpers",
|
|
"@xla///stream_executor/cuda:cudart_stub",
|
|
"@xla///stream_executor/cuda:cusolver_lib",
|
|
"@com_google_absl//absl/container:flat_hash_map",
|
|
"@com_google_absl//absl/strings:str_format",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
"@pybind11",
|
|
],
|
|
)
|
|
|
|
cc_library(
|
|
name = "cusparse_kernels",
|
|
srcs = ["//jaxlib/gpu:sparse_kernels.cc"],
|
|
hdrs = ["//jaxlib/gpu:sparse_kernels.h"],
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_vendor",
|
|
"//jaxlib:handle_pool",
|
|
"//jaxlib:kernel_helpers",
|
|
"@xla///service:custom_call_status",
|
|
"@xla///stream_executor/cuda:cudart_stub",
|
|
"@xla///stream_executor/cuda:cusparse_lib",
|
|
"@com_google_absl//absl/status",
|
|
"@com_google_absl//absl/status:statusor",
|
|
"@com_google_absl//absl/synchronization",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
],
|
|
)
|
|
|
|
pybind_extension(
|
|
name = "_sparse",
|
|
srcs = ["//jaxlib/gpu:sparse.cc"],
|
|
copts = [
|
|
"-fexceptions",
|
|
"-fno-strict-aliasing",
|
|
],
|
|
features = ["-use_header_modules"],
|
|
module_name = "_sparse",
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_vendor",
|
|
":cusparse_kernels",
|
|
"//jaxlib:kernel_pybind11_helpers",
|
|
"@xla///stream_executor/cuda:cudart_stub",
|
|
"@xla///stream_executor/cuda:cusparse_lib",
|
|
"@com_google_absl//absl/algorithm:container",
|
|
"@com_google_absl//absl/base",
|
|
"@com_google_absl//absl/base:core_headers",
|
|
"@com_google_absl//absl/container:flat_hash_map",
|
|
"@com_google_absl//absl/hash",
|
|
"@com_google_absl//absl/memory",
|
|
"@com_google_absl//absl/strings",
|
|
"@com_google_absl//absl/strings:str_format",
|
|
"@com_google_absl//absl/synchronization",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
"@pybind11",
|
|
],
|
|
)
|
|
|
|
cc_library(
|
|
name = "cuda_lu_pivot_kernels",
|
|
srcs = [
|
|
"//jaxlib/gpu:lu_pivot_kernels.cc",
|
|
],
|
|
hdrs = ["//jaxlib/gpu:lu_pivot_kernels.h"],
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_lu_pivot_kernels_impl",
|
|
":cuda_vendor",
|
|
"//jaxlib:kernel_helpers",
|
|
"@xla///service:custom_call_status",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
],
|
|
)
|
|
|
|
cuda_library(
|
|
name = "cuda_lu_pivot_kernels_impl",
|
|
srcs = [
|
|
"//jaxlib/gpu:lu_pivot_kernels.cu.cc",
|
|
],
|
|
hdrs = ["//jaxlib/gpu:lu_pivot_kernels.h"],
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_vendor",
|
|
"//jaxlib:kernel_helpers",
|
|
"@xla///service:custom_call_status",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
],
|
|
)
|
|
|
|
pybind_extension(
|
|
name = "_linalg",
|
|
srcs = ["//jaxlib/gpu:linalg.cc"],
|
|
copts = [
|
|
"-fexceptions",
|
|
"-fno-strict-aliasing",
|
|
],
|
|
features = ["-use_header_modules"],
|
|
module_name = "_linalg",
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_lu_pivot_kernels",
|
|
":cuda_lu_pivot_kernels_impl",
|
|
":cuda_vendor",
|
|
"//jaxlib:kernel_pybind11_helpers",
|
|
"@xla///stream_executor/cuda:cudart_stub",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
"@pybind11",
|
|
],
|
|
)
|
|
|
|
cc_library(
|
|
name = "cuda_prng_kernels",
|
|
srcs = [
|
|
"//jaxlib/gpu:prng_kernels.cc",
|
|
],
|
|
hdrs = ["//jaxlib/gpu:prng_kernels.h"],
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_prng_kernels_impl",
|
|
":cuda_vendor",
|
|
"//jaxlib:kernel_helpers",
|
|
"@xla///service:custom_call_status",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
],
|
|
)
|
|
|
|
cuda_library(
|
|
name = "cuda_prng_kernels_impl",
|
|
srcs = [
|
|
"//jaxlib/gpu:prng_kernels.cu.cc",
|
|
],
|
|
hdrs = ["//jaxlib/gpu:prng_kernels.h"],
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_vendor",
|
|
"//jaxlib:kernel_helpers",
|
|
"@xla///service:custom_call_status",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
],
|
|
)
|
|
|
|
pybind_extension(
|
|
name = "_prng",
|
|
srcs = ["//jaxlib/gpu:prng.cc"],
|
|
copts = [
|
|
"-fexceptions",
|
|
"-fno-strict-aliasing",
|
|
],
|
|
features = ["-use_header_modules"],
|
|
module_name = "_prng",
|
|
deps = [
|
|
":cuda_gpu_kernel_helpers",
|
|
":cuda_prng_kernels",
|
|
"//jaxlib:kernel_pybind11_helpers",
|
|
"@xla///stream_executor/cuda:cudart_stub",
|
|
"@local_config_cuda//cuda:cuda_headers",
|
|
"@pybind11",
|
|
],
|
|
)
|
|
|
|
cc_library(
|
|
name = "cuda_gpu_kernels",
|
|
srcs = ["//jaxlib/gpu:gpu_kernels.cc"],
|
|
visibility = ["//visibility:public"],
|
|
deps = [
|
|
":cublas_kernels",
|
|
":cuda_lu_pivot_kernels",
|
|
":cuda_prng_kernels",
|
|
":cuda_vendor",
|
|
":cusolver_kernels",
|
|
":cusparse_kernels",
|
|
"@xla///service:custom_call_target_registry",
|
|
],
|
|
alwayslink = 1,
|
|
)
|
|
|
|
py_library(
|
|
name = "cuda_gpu_support",
|
|
deps = [
|
|
":_blas",
|
|
":_linalg",
|
|
":_prng",
|
|
":_rnn",
|
|
":_solver",
|
|
":_sparse",
|
|
],
|
|
)
|
|
|
|
# We cannot nest select and if_cuda_is_configured so we introduce
|
|
# a standalone py_library target.
|
|
py_library(
|
|
name = "gpu_only_test_deps",
|
|
# `if_cuda_is_configured` will default to `[]`.
|
|
deps = if_cuda_is_configured([":cuda_gpu_support"]),
|
|
)
|