llvm
/
torch-mlir
mirror of https://github.com/llvm/torch-mlir
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
torch-mlir/tools/mnist-playground/CMakeLists.txt

49 lines
1.1 KiB
CMake
Raw Normal View History

Add hopefully short-lived mnist-playground utility. This unblocks backend progress while the PyTorch frontend work is coming online. Hopefully we can delete this soon. See tools/mnist-playground/README.md for more context on what this tool is for, next steps, and current status.
2020-10-03 05:57:35 +08:00
# TODO: This is copied from frontends/pytorch/csrc/c10_dispatch/CMakeLists.txt
# What is the idiomatic way of sharing this in CMake?
include_directories(
${TORCH_INCLUDE_DIRS}
${TORCH_INSTALL_PREFIX}/include/TH
${TORCH_INSTALL_PREFIX}/include/THC/opt/pytorch/pytorch
${CMAKE_CURRENT_SOURCE_DIR}
${CMAKE_CURRENT_BINARY_DIR}
${PYTHON_INCLUDE_DIRS}
)
link_directories("${TORCH_INSTALL_PREFIX}/lib")
set(LLVM_LINK_COMPONENTS
Core
Support
nativecodegen
)
Start reworking towards a shared library build. * Need to have a dag of shared library deps in order to interop across python extensions (as presented in ODM). * Introduced add_npcomp_library and friends to mirror the MLIR setup. * Adds a libNPCOMP.so shared library. * Redirects tools and extensions to link against libNPCOMP.so (instead of static libs). * Moves all libraries to lib/, all binaries to bin/ and all python extensions to python/. The invariant is that the rpaths are setup to have a one level directory structure. * Reworks the _torch_mlir extension to build like the others (still need to come up with a consolidated rule to do this instead of open coded). * Includes an upstream version bump to pick up needed changes. Sizes with dynamic linking (stripped, release, asserts enabled): libNPCOMP.so: 43M (includes much of the underlying LLVM codegen deps) libMLIR.so: 31M _npcomp.so: 1.6M (python extension) _torch_mlir.so: 670K (python extension) npcomp-capi-ir-test: 6.3K npcomp-opt: 351K npcomp-run-mlir: 461K mnist-playground: 530K Still more can be done to normalize and optimize but this gets us structurally to the starting point.
2020-10-09 09:29:59 +08:00
add_npcomp_executable(mnist-playground
Add hopefully short-lived mnist-playground utility. This unblocks backend progress while the PyTorch frontend work is coming online. Hopefully we can delete this soon. See tools/mnist-playground/README.md for more context on what this tool is for, next steps, and current status.
2020-10-03 05:57:35 +08:00
mnist-playground.cpp
)
llvm_update_compile_flags(mnist-playground)
get_property(dialect_libs GLOBAL PROPERTY MLIR_DIALECT_LIBS)
get_property(conversion_libs GLOBAL PROPERTY MLIR_CONVERSION_LIBS)
target_link_libraries(mnist-playground PRIVATE
Start reworking towards a shared library build. * Need to have a dag of shared library deps in order to interop across python extensions (as presented in ODM). * Introduced add_npcomp_library and friends to mirror the MLIR setup. * Adds a libNPCOMP.so shared library. * Redirects tools and extensions to link against libNPCOMP.so (instead of static libs). * Moves all libraries to lib/, all binaries to bin/ and all python extensions to python/. The invariant is that the rpaths are setup to have a one level directory structure. * Reworks the _torch_mlir extension to build like the others (still need to come up with a consolidated rule to do this instead of open coded). * Includes an upstream version bump to pick up needed changes. Sizes with dynamic linking (stripped, release, asserts enabled): libNPCOMP.so: 43M (includes much of the underlying LLVM codegen deps) libMLIR.so: 31M _npcomp.so: 1.6M (python extension) _torch_mlir.so: 670K (python extension) npcomp-capi-ir-test: 6.3K npcomp-opt: 351K npcomp-run-mlir: 461K mnist-playground: 530K Still more can be done to normalize and optimize but this gets us structurally to the starting point.
2020-10-09 09:29:59 +08:00
# Shared library deps first ensure we get most of what we need from libraries.
NPCOMP
MLIR
Add hopefully short-lived mnist-playground utility. This unblocks backend progress while the PyTorch frontend work is coming online. Hopefully we can delete this soon. See tools/mnist-playground/README.md for more context on what this tool is for, next steps, and current status.
2020-10-03 05:57:35 +08:00
MLIRAnalysis
MLIREDSC
MLIRExecutionEngine
MLIRIR
MLIRJitRunner
MLIRLLVMIR
MLIRParser
MLIRTargetLLVMIR
MLIRSupport
NPCOMPInitAll
[RefBackend] Move runtime related code under npcomp/RefBackend/ Other than the dialect definitions (which will live in standard Dialect/ subdirectory), the goal here is to keep RefBackend-related code nested in {include/npcomp,lib,test}/RefBackend.
2020-10-08 07:11:41 +08:00
NPCOMPRefBackendJITHelpers
Add hopefully short-lived mnist-playground utility. This unblocks backend progress while the PyTorch frontend work is coming online. Hopefully we can delete this soon. See tools/mnist-playground/README.md for more context on what this tool is for, next steps, and current status.
2020-10-03 05:57:35 +08:00
${conversion_libs}
${dialect_libs}
${TORCH_LIBRARIES}
)
add_dependencies(mnist-playground
NPCOMPCompilerRuntimeShlib
)