pytorch
diff --git a/‎CMakeLists.txt
Lines changed: 11 additions & 0 deletions b/‎CMakeLists.txt
Lines changed: 11 additions & 0 deletions
diff --git a/‎torch/_inductor/codecache.py
Lines changed: 15 additions & 0 deletions b/‎torch/_inductor/codecache.py
Lines changed: 15 additions & 0 deletions
@@ -349,6 +349,8 @@ cmake_dependent_option(
     "NOT INTERN_BUILD_MOBILE" OFF)
 cmake_dependent_option(
     BUILD_FUNCTORCH "Build Functorch" ON "BUILD_PYTHON" OFF)
+cmake_dependent_option(
+    BUILD_BUNDLE_PTXAS "Bundle PTX into torch/bin fodler" OFF "USE_CUDA" OFF)
 
 option(USE_MIMALLOC "Use mimalloc" OFF)
 # Enable third party mimalloc library to improve memory allocation performance on Windows.
@@ -1230,3 +1232,12 @@ if(DEFINED USE_CUSTOM_DEBINFO)
     set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -g")
     set(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -g")
 endif()
+
+# Bundle PTXAS if needed
+if(BUILD_BUNDLE_PTXAS AND USE_CUDA)
+   if(NOT EXISTS "${PROJECT_SOURCE_DIR}/build/bin/ptxas")
+     message(STATUS "Copying PTXAS into the bin folder")
+     file(COPY "${CUDAToolkit_BIN_DIR}/ptxas" DESTINATION "${PROJECT_BINARY_DIR}")
+   endif()
+   install(PROGRAMS "${PROJECT_BINARY_DIR}/ptxas" DESTINATION "${CMAKE_INSTALL_BINDIR}")
+endif()
@@ -2277,6 +2277,20 @@ def caching_device_properties():
             device_interface.Worker.get_device_properties()
 
 
+def _set_triton_ptxas_path() -> None:
+    if os.environ.get("TRITON_PTXAS_PATH") is not None:
+        return
+    ptxas_path = os.path.abspath(
+        os.path.join(os.path.dirname(__file__), "..", "bin", "ptxas")
+    )
+    if not os.path.exists(ptxas_path):
+        return
+    if os.path.isfile(ptxas_path) and os.access(ptxas_path, os.X_OK):
+        os.environ["TRITON_PTXAS_PATH"] = ptxas_path
+    else:
+        warnings.warn(f"{ptxas_path} exists but is not an executable")
+
+
 def _worker_compile(
     kernel_name: str, source_code: str, cc: int, device: torch.device
 ) -> None:
@@ -2287,6 +2301,7 @@ def _worker_compile(
 
 
 def _load_kernel(kernel_name: str, source_code: str) -> ModuleType:
+    _set_triton_ptxas_path()
     kernel = TritonCodeCache.load(kernel_name, source_code)
     kernel.precompile()
     return kernel