Set USE_KINETO=1 (#49201)

Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49201

This unblocks kineto profiler for 1.8 release.
This PR supercedes https://github.com/pytorch/pytorch/pull/48391
Note: this will somewhat increase the size of linux server binaries, bc
we add libkineto.a and libcupti_static.a:
-rw-r--r-- 1 jenkins jenkins 1107502 Dec 10 21:16 build/lib/libkineto.a
-rw-r--r-- 1 root root 13699658 Nov 13  2019 /usr/local/cuda/lib64/libcupti_static.a

Test Plan:
CI
https://github.com/pytorch/pytorch/pull/48391

Imported from OSS

Reviewed By: ngimel

Differential Revision: D25480770

fbshipit-source-id: 037cd774f5547d9918d6055ef5cc952a54e48e4c

CMakeLists.txt

@@ -159,7 +159,8 @@ cmake_dependent_option(
     USE_STATIC_CUDNN "Use cuDNN static libraries" OFF
     "USE_CUDNN" OFF)
 option(USE_FBGEMM "Use FBGEMM (quantized 8-bit server operators)" ON)
-option(USE_KINETO "Use Kineto profiling library" OFF)
+option(USE_KINETO "Use Kineto profiling library" ON)
+option(USE_CUPTI_SO "Use CUPTI as a shared library" OFF)
 option(USE_FAKELOWP "Use FakeLowp operators" OFF)
 option(USE_FFMPEG "Use ffmpeg" OFF)
 option(USE_GFLAGS "Use GFLAGS" OFF)
@@ -515,31 +516,12 @@ if(USE_FBGEMM AND ((CMAKE_SYSTEM_PROCESSOR STREQUAL "x86_64" AND CMAKE_SIZEOF_VO
   set(USE_FBGEMM OFF)
 endif()
 
-if(USE_KINETO AND INTERN_BUILD_MOBILE)
-  message(STATUS "Not using libkineto in a mobile build.")
-  set(USE_KINETO OFF)
-endif()
-
-if(USE_KINETO AND (NOT USE_CUDA))
-  message(STATUS "Not using libkineto in a non-CUDA build.")
-  set(USE_KINETO OFF)
-endif()
-
-if(USE_KINETO AND MSVC)
-  message(STATUS "Not using libkineto in a Windows build.")
-  set(USE_KINETO OFF)
-endif()
-
 include(cmake/Dependencies.cmake)
 
 if(USE_FBGEMM)
   string(APPEND CMAKE_CXX_FLAGS " -DUSE_FBGEMM")
 endif()
 
-if(USE_KINETO)
-  string(APPEND CMAKE_CXX_FLAGS " -DUSE_KINETO")
-endif()
-
 if(USE_QNNPACK)
   string(APPEND CMAKE_CXX_FLAGS " -DUSE_QNNPACK")
 endif()

cmake/Dependencies.cmake

@@ -1788,8 +1788,21 @@ list(APPEND Caffe2_DEPENDENCY_LIBS fmt::fmt-header-only)
 set(BUILD_SHARED_LIBS ${TEMP_BUILD_SHARED_LIBS} CACHE BOOL "Build shared libs" FORCE)
 
 # ---[ Kineto
+if(USE_KINETO AND INTERN_BUILD_MOBILE)
+  message(STATUS "Not using libkineto in a mobile build.")
+  set(USE_KINETO OFF)
+endif()
+
+if(USE_KINETO AND (NOT USE_CUDA))
+  message(STATUS "Not using libkineto in a non-CUDA build.")
+  set(USE_KINETO OFF)
+endif()
+
+if(USE_KINETO AND MSVC)
+  message(STATUS "Not using libkineto in a Windows build.")
+  set(USE_KINETO OFF)
+endif()
 if(USE_KINETO)
-  if(USE_KINETO AND NOT TARGET kineto)
   set(CAFFE2_THIRD_PARTY_ROOT "${PROJECT_SOURCE_DIR}/third_party" CACHE STRING "")
   set(KINETO_SOURCE_DIR "${CAFFE2_THIRD_PARTY_ROOT}/kineto/libkineto" CACHE STRING "")
   set(KINETO_BUILD_TESTS OFF CACHE BOOL "")
@@ -1802,29 +1815,43 @@ if(USE_KINETO)
   message(STATUS "  KINETO_LIBRARY_TYPE = ${KINETO_LIBRARY_TYPE}")
   message(STATUS "  CUDA_SOURCE_DIR = ${CUDA_SOURCE_DIR}")
 
+  if(EXISTS ${CUDA_SOURCE_DIR}/extras/CUPTI/lib64/libcupti_static.a)
+    set(CUDA_cupti_LIBRARY "${CUDA_SOURCE_DIR}/extras/CUPTI/lib64/libcupti_static.a")
+  elseif(EXISTS ${CUDA_SOURCE_DIR}/lib64/libcupti_static.a)
+    set(CUDA_cupti_LIBRARY "${CUDA_SOURCE_DIR}/lib64/libcupti_static.a")
+  elseif(USE_CUPTI_SO)
+    if(EXISTS ${CUDA_SOURCE_DIR}/extras/CUPTI/lib64/libcupti.so)
+      set(CUDA_cupti_LIBRARY "${CUDA_SOURCE_DIR}/extras/CUPTI/lib64/libcupti.so")
+    elseif(EXISTS ${CUDA_SOURCE_DIR}/lib64/libcupti.so)
+      set(CUDA_cupti_LIBRARY "${CUDA_SOURCE_DIR}/lib64/libcupti.so")
+    endif()
+  endif()
+
   if(EXISTS ${CUDA_SOURCE_DIR}/extras/CUPTI/include)
     set(CUPTI_INCLUDE_DIR "${CUDA_SOURCE_DIR}/extras/CUPTI/include")
   elseif(EXISTS ${CUDA_SOURCE_DIR}/include/cupti.h)
     set(CUPTI_INCLUDE_DIR "${CUDA_SOURCE_DIR}/include")
   endif()
 
-    if((NOT DEFINED CUDA_cupti_LIBRARY) OR (${CUDA_cupti_LIBRARY} STREQUAL "CUDA_cupti_LIBRARY-NOTFOUND"))
+  set(FOUND_CUPTI FALSE)
-      if(EXISTS ${CUDA_SOURCE_DIR}/extras/CUPTI/lib64/libcupti_static.a)
+  if((DEFINED CUPTI_INCLUDE_DIR) AND (DEFINED CUDA_cupti_LIBRARY))
-        set(CUDA_cupti_LIBRARY "${CUDA_SOURCE_DIR}/extras/CUPTI/lib64/libcupti_static.a")
+    if((CUDA_cupti_LIBRARY MATCHES "libcupti_static.a") OR ((CUDA_cupti_LIBRARY MATCHES "libcupti.so") AND (USE_CUPTI_SO)))
-      elseif(EXISTS ${CUDA_SOURCE_DIR}/lib64/libcupti_static.a)
+      set(FOUND_CUPTI TRUE)
-        set(CUDA_cupti_LIBRARY "${CUDA_SOURCE_DIR}/lib64/libcupti_static.a")
-      elseif(EXISTS ${CUDA_SOURCE_DIR}/extras/CUPTI/lib64/libcupti.so)
-        set(CUDA_cupti_LIBRARY "${CUDA_SOURCE_DIR}/extras/CUPTI/lib64/libcupti.so")
-      elseif(EXISTS ${CUDA_SOURCE_DIR}/lib64/libcupti.so)
-        set(CUDA_cupti_LIBRARY "${CUDA_SOURCE_DIR}/lib64/libcupti.so")
     endif()
   endif()
+
+  if(FOUND_CUPTI)
     message(STATUS "  CUDA_cupti_LIBRARY = ${CUDA_cupti_LIBRARY}")
     message(STATUS "  CUPTI_INCLUDE_DIR = ${CUPTI_INCLUDE_DIR}")
-
+    if(NOT TARGET kineto)
       add_subdirectory("${KINETO_SOURCE_DIR}")
       message(STATUS "Configured Kineto as a dependency.")
     endif()
 
+    string(APPEND CMAKE_CXX_FLAGS " -DUSE_KINETO")
     list(APPEND Caffe2_DEPENDENCY_LIBS kineto)
+  else()
+    message(STATUS "Could not find CUPTI library, skipping Kineto build")
+    set(USE_KINETO OFF)
+  endif()
 endif()

setup.py

@@ -33,8 +33,8 @@
 #   USE_FBGEMM=0
 #     disables the FBGEMM build
 #
-#   USE_KINETO=1
+#   USE_KINETO=0
-#     enables experimental usage of libkineto
+#     disables usage of libkineto library for profiling
 #
 #   USE_NUMPY=0
 #     disables the NumPy build

torch/csrc/autograd/profiler_kineto.cpp

@@ -360,7 +360,7 @@ void ProfilerResult::save(const std::string& path) {
 
 bool kinetoAvailable() {
 #ifdef USE_KINETO
-  return true;
+  return at::hasCUDA();
 #else
   return false;
 #endif