llvm
/
torch-mlir
mirror of https://github.com/llvm/torch-mlir
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
torch-mlir/externals/llvm-external-projects/torch-mlir-dialects/CMakeLists.txt

65 lines
2.6 KiB
CMake
Raw Normal View History

Add TMTensor dialect to torch-mlir This is intended to explore support for non-structured ops that can't be modeled by Linalg dialect. `tm_tensor.scan` and `tm_tensor.scatter` are added as the first such ops. The dialect should aim to be upstreamed in the future.
2022-02-03 07:01:38 +08:00
if(CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR)
Fix out-of-tree build of torch-mlir-dialects (#726) Follows up on #623 for out-of-tree builds of torch-mlir, which added building `torch-mir-dialects` as a subdirectory. Our goal is to support both in-tree and out-of-tree builds of `torch-mlir` with minimum hassle, for instance by using the same variable names in both setups. Specific changes to `externals/llvm-external-projects/torch-mlir-dialects/CMakeLists.txt`: - We use `MLIR_FOUND` to detect that it is being build as a subdirectory and the llvm+mlir cmake infrastructure is already set up (via find_package in the parent build) as opposed to an in-tree build. - For in-tree, the setting of variables and loading of llvm+mlir cmake infrastructure is now conditionally performed. - For in-tree, the names of cmake variables being defined for are adjusted to match those `llvm-project` makes available through `find_package(MLIR REQUIRED CONFIG)`, under the assumption that those are the more "standardized" names. Co-authored-by: Clément Fournier <clement.fournier@amd.com> Co-authored-by: Liam Fitzpatrick <liam.fitzpatrick@xilinx.com>
2022-04-04 17:37:28 +08:00
# The cmake configuration to build torch-mlir-dialects as
# an out-of-tree project has not been implemented. It can
# be built as part of LLVM or as a subdirectory of torch-mlir.
message(FATAL_ERROR
"This project is intended to be built as part of LLVM via "
"-DLLVM_EXTERNAL_PROJECTS=torch-mlir-dialects "
"-DLLVM_EXTERNAL_TORCH_MLIR_DIALECTS_SOURCE_DIR=${CMAKE_CURRENT_SOURCE_DIR}")
Add TMTensor dialect to torch-mlir This is intended to explore support for non-structured ops that can't be modeled by Linalg dialect. `tm_tensor.scan` and `tm_tensor.scatter` are added as the first such ops. The dialect should aim to be upstreamed in the future.
2022-02-03 07:01:38 +08:00
endif()
option(MLIR_ENABLE_BINDINGS_PYTHON "Enables MLIR Python Bindings" OFF)
set(TORCH_MLIR_DIALECTS_SOURCE_DIR "${CMAKE_CURRENT_SOURCE_DIR}")
set(TORCH_MLIR_DIALECTS_BINARY_DIR "${CMAKE_CURRENT_BINARY_DIR}")
message(STATUS "Building torch-mlir-dialects project at ${TORCH_MLIR_DIALECTS_SOURCE_DIR} (into ${TORCH_MLIR_DIALECTS_BINARY_DIR})")
Fix out-of-tree build of torch-mlir-dialects (#726) Follows up on #623 for out-of-tree builds of torch-mlir, which added building `torch-mir-dialects` as a subdirectory. Our goal is to support both in-tree and out-of-tree builds of `torch-mlir` with minimum hassle, for instance by using the same variable names in both setups. Specific changes to `externals/llvm-external-projects/torch-mlir-dialects/CMakeLists.txt`: - We use `MLIR_FOUND` to detect that it is being build as a subdirectory and the llvm+mlir cmake infrastructure is already set up (via find_package in the parent build) as opposed to an in-tree build. - For in-tree, the setting of variables and loading of llvm+mlir cmake infrastructure is now conditionally performed. - For in-tree, the names of cmake variables being defined for are adjusted to match those `llvm-project` makes available through `find_package(MLIR REQUIRED CONFIG)`, under the assumption that those are the more "standardized" names. Co-authored-by: Clément Fournier <clement.fournier@amd.com> Co-authored-by: Liam Fitzpatrick <liam.fitzpatrick@xilinx.com>
2022-04-04 17:37:28 +08:00
if(MLIR_FOUND)
message(STATUS "LLVM and MLIR packages have already been configured.")
else()
message(STATUS "torch-mlir-dialect is being built in-tree")
# An in-tree build, LLVM hasn't been installed yet.
# We compute these properties manually, they're otherwise
# contributed by find_package(MLIR...)
set(MLIR_MAIN_SRC_DIR ${LLVM_MAIN_SRC_DIR}/../mlir)
set(MLIR_INCLUDE_DIR ${LLVM_MAIN_SRC_DIR}/../mlir/include)
set(MLIR_GENERATED_INCLUDE_DIR ${LLVM_BINARY_DIR}/tools/mlir/include)
set(MLIR_INCLUDE_DIRS ${MLIR_INCLUDE_DIR} ${MLIR_GENERATED_INCLUDE_DIR})
set(LLVM_INCLUDE_DIRS ${LLVM_MAIN_INCLUDE_DIR})
Add TMTensor dialect to torch-mlir This is intended to explore support for non-structured ops that can't be modeled by Linalg dialect. `tm_tensor.scan` and `tm_tensor.scatter` are added as the first such ops. The dialect should aim to be upstreamed in the future.
2022-02-03 07:01:38 +08:00
Fix out-of-tree build of torch-mlir-dialects (#726) Follows up on #623 for out-of-tree builds of torch-mlir, which added building `torch-mir-dialects` as a subdirectory. Our goal is to support both in-tree and out-of-tree builds of `torch-mlir` with minimum hassle, for instance by using the same variable names in both setups. Specific changes to `externals/llvm-external-projects/torch-mlir-dialects/CMakeLists.txt`: - We use `MLIR_FOUND` to detect that it is being build as a subdirectory and the llvm+mlir cmake infrastructure is already set up (via find_package in the parent build) as opposed to an in-tree build. - For in-tree, the setting of variables and loading of llvm+mlir cmake infrastructure is now conditionally performed. - For in-tree, the names of cmake variables being defined for are adjusted to match those `llvm-project` makes available through `find_package(MLIR REQUIRED CONFIG)`, under the assumption that those are the more "standardized" names. Co-authored-by: Clément Fournier <clement.fournier@amd.com> Co-authored-by: Liam Fitzpatrick <liam.fitzpatrick@xilinx.com>
2022-04-04 17:37:28 +08:00
# Configure CMake and tablegen.
list(APPEND CMAKE_MODULE_PATH ${MLIR_MAIN_SRC_DIR}/cmake/modules)
list(APPEND CMAKE_MODULE_PATH ${LLVM_MAIN_SRC_DIR}/cmake)
include(TableGen)
include(AddLLVM)
include(AddMLIR)
endif()
Add TMTensor dialect to torch-mlir This is intended to explore support for non-structured ops that can't be modeled by Linalg dialect. `tm_tensor.scan` and `tm_tensor.scatter` are added as the first such ops. The dialect should aim to be upstreamed in the future.
2022-02-03 07:01:38 +08:00
function(torch_mlir_dialects_target_includes target)
set(_dirs
Fix out-of-tree build of torch-mlir-dialects (#726) Follows up on #623 for out-of-tree builds of torch-mlir, which added building `torch-mir-dialects` as a subdirectory. Our goal is to support both in-tree and out-of-tree builds of `torch-mlir` with minimum hassle, for instance by using the same variable names in both setups. Specific changes to `externals/llvm-external-projects/torch-mlir-dialects/CMakeLists.txt`: - We use `MLIR_FOUND` to detect that it is being build as a subdirectory and the llvm+mlir cmake infrastructure is already set up (via find_package in the parent build) as opposed to an in-tree build. - For in-tree, the setting of variables and loading of llvm+mlir cmake infrastructure is now conditionally performed. - For in-tree, the names of cmake variables being defined for are adjusted to match those `llvm-project` makes available through `find_package(MLIR REQUIRED CONFIG)`, under the assumption that those are the more "standardized" names. Co-authored-by: Clément Fournier <clement.fournier@amd.com> Co-authored-by: Liam Fitzpatrick <liam.fitzpatrick@xilinx.com>
2022-04-04 17:37:28 +08:00
$<BUILD_INTERFACE:${MLIR_INCLUDE_DIRS}>
Add TMTensor dialect to torch-mlir This is intended to explore support for non-structured ops that can't be modeled by Linalg dialect. `tm_tensor.scan` and `tm_tensor.scatter` are added as the first such ops. The dialect should aim to be upstreamed in the future.
2022-02-03 07:01:38 +08:00
$<BUILD_INTERFACE:${TORCH_MLIR_DIALECTS_SOURCE_DIR}/include>
$<BUILD_INTERFACE:${TORCH_MLIR_DIALECTS_BINARY_DIR}/include>
)
# In LLVM parlance, the actual target may just be an interface and may not
# be responsible for actually compiling anything. The corresponding obj.
# target, when present, is just used for compilation and does not
# contribute to the interface properties.
# TODO: Normalize this upstream.
Missing ";" bewtween includes in MHLOTargets.cmake (#2074) Co-authored-by: Tilmann Bartsch <info@tebartsch.ai>
2023-04-30 07:05:20 +08:00
target_include_directories(${target} PUBLIC "${_dirs}")
Add TMTensor dialect to torch-mlir This is intended to explore support for non-structured ops that can't be modeled by Linalg dialect. `tm_tensor.scan` and `tm_tensor.scatter` are added as the first such ops. The dialect should aim to be upstreamed in the future.
2022-02-03 07:01:38 +08:00
if(TARGET obj.${target})
Missing ";" bewtween includes in MHLOTargets.cmake (#2074) Co-authored-by: Tilmann Bartsch <info@tebartsch.ai>
2023-04-30 07:05:20 +08:00
target_include_directories(obj.${target} PRIVATE "${_dirs}")
Add TMTensor dialect to torch-mlir This is intended to explore support for non-structured ops that can't be modeled by Linalg dialect. `tm_tensor.scan` and `tm_tensor.scatter` are added as the first such ops. The dialect should aim to be upstreamed in the future.
2022-02-03 07:01:38 +08:00
endif()
endfunction()
Fix out-of-tree build of torch-mlir-dialects (#726) Follows up on #623 for out-of-tree builds of torch-mlir, which added building `torch-mir-dialects` as a subdirectory. Our goal is to support both in-tree and out-of-tree builds of `torch-mlir` with minimum hassle, for instance by using the same variable names in both setups. Specific changes to `externals/llvm-external-projects/torch-mlir-dialects/CMakeLists.txt`: - We use `MLIR_FOUND` to detect that it is being build as a subdirectory and the llvm+mlir cmake infrastructure is already set up (via find_package in the parent build) as opposed to an in-tree build. - For in-tree, the setting of variables and loading of llvm+mlir cmake infrastructure is now conditionally performed. - For in-tree, the names of cmake variables being defined for are adjusted to match those `llvm-project` makes available through `find_package(MLIR REQUIRED CONFIG)`, under the assumption that those are the more "standardized" names. Co-authored-by: Clément Fournier <clement.fournier@amd.com> Co-authored-by: Liam Fitzpatrick <liam.fitzpatrick@xilinx.com>
2022-04-04 17:37:28 +08:00
# TODO: Needed for tablegen. Remove.
include_directories(SYSTEM ${MLIR_INCLUDE_DIRS})
include_directories(SYSTEM ${LLVM_INCLUDE_DIRS})
include_directories(SYSTEM ${TORCH_MLIR_DIALECTS_SOURCE_DIR}/include)
Add TMTensor dialect to torch-mlir This is intended to explore support for non-structured ops that can't be modeled by Linalg dialect. `tm_tensor.scan` and `tm_tensor.scatter` are added as the first such ops. The dialect should aim to be upstreamed in the future.
2022-02-03 07:01:38 +08:00
add_subdirectory(include)
add_subdirectory(lib)
add_subdirectory(tools)
add_subdirectory(test)