Skip to content

Fix Windows CUDA 12.9 SM120 compilation #32114

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Hariharan Seshadri (hariharans29) merged 2 commits into from
Aug 24, 2026
Merged

Fix Windows CUDA 12.9 SM120 compilation #32114

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e7f4e91
Initial plan
Copilot Aug 16, 2026
118890c
Fix CUDA 12.9 SM120 build on Windows
Copilot Aug 16, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 20 additions & 0 deletions cmake/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1516,6 +1516,26 @@ if (onnxruntime_USE_CUDA)
endif()
find_package(CUDAToolkit REQUIRED)

if(MSVC AND CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 12.9
AND CMAKE_CUDA_COMPILER_VERSION VERSION_LESS 13.0)
foreach(_cuda_include_dir IN LISTS CUDAToolkit_INCLUDE_DIRS)
set(_clusterlaunchcontrol_header
"${_cuda_include_dir}/cuda/__ptx/instructions/generated/clusterlaunchcontrol.h")
if(EXISTS "${_clusterlaunchcontrol_header}")
# CUDA 12.9 uses Windows's 32-bit long for PTX 64-bit operands. Shadow the toolkit
# header with the longlong2 correction shipped by newer CCCL releases.
set(_cuda_12_9_fix_dir "${CMAKE_CURRENT_BINARY_DIR}/cuda_12_9_fix")
ort_patch_cuda_12_9_clusterlaunchcontrol_header(
"${_clusterlaunchcontrol_header}"
"${_cuda_12_9_fix_dir}/cuda/__ptx/instructions/generated/clusterlaunchcontrol.h")
if(EXISTS "${_cuda_12_9_fix_dir}/cuda/__ptx/instructions/generated/clusterlaunchcontrol.h")
include_directories(BEFORE "${_cuda_12_9_fix_dir}")
endif()
break()
endif()
endforeach()
endif()

# Note: The minimum required CUDA version is greater than 11.8.
add_definitions("-DENABLE_BF16")
message(STATUS "CUDA Toolkit version is greater or equal than 11.8, enable -DENABLE_BF16 flag")
Expand Down
17 changes: 17 additions & 0 deletions cmake/external/cuda_configuration.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -73,6 +73,23 @@ macro(setup_cuda_compiler)
endif()
endmacro()

function(ort_patch_cuda_12_9_clusterlaunchcontrol_header src dst)
if(NOT EXISTS "${src}")
return()
endif()

file(READ "${src}" _content)
set(_orig "${_content}")
string(REPLACE "reinterpret_cast<long2*>" "reinterpret_cast<longlong2*>" _content "${_content}")
if(NOT _content STREQUAL _orig)
get_filename_component(_dst_dir "${dst}" DIRECTORY)
file(MAKE_DIRECTORY "${_dst_dir}")
file(WRITE "${dst}" "${_content}")
elseif(EXISTS "${dst}")
file(REMOVE "${dst}")
endif()
endfunction()

macro(setup_cuda_architectures)
# cmake-format: off
# Initialize and normalize CMAKE_CUDA_ARCHITECTURES before enabling CUDA.
Expand Down
Loading