# Copyright (c) 2022-2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
#
# See LICENSE for license information.

cmake_minimum_required(VERSION 3.18)

if(NOT DEFINED CMAKE_CUDA_ARCHITECTURES)
  set(CMAKE_CUDA_ARCHITECTURES 70 80 89 90)
endif()

set(CMAKE_CXX_STANDARD 17)
set(CMAKE_CUDA_STANDARD 17)
set(CMAKE_CUDA_STANDARD_REQUIRED ON)

project(transformer_engine LANGUAGES CUDA CXX)

set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} --threads 4")
if (CMAKE_BUILD_TYPE STREQUAL "Debug")
  set(CMAKE_CUDA_FLAGS_DEBUG "${CMAKE_CUDA_FLAGS_DEBUG} -G")
endif()

find_package(CUDAToolkit REQUIRED cublas nvToolsExt)

# Check for cuDNN frontend API
set(CUDNN_FRONTEND_INCLUDE_DIR
    "${CMAKE_SOURCE_DIR}/../3rdparty/cudnn-frontend/include")
if(NOT EXISTS "${CUDNN_FRONTEND_INCLUDE_DIR}")
    message(FATAL_ERROR
            "Could not find cuDNN frontend API. "
            "Try running 'git submodule update --init --recursive' "
            "within the Transformer Engine source.")
endif()
include(${CMAKE_SOURCE_DIR}/../3rdparty/cudnn-frontend/cmake/cuDNN.cmake)

find_package(Python COMPONENTS Interpreter Development.Module REQUIRED)

include_directories(${PROJECT_SOURCE_DIR})

add_subdirectory(common)
if(NVTE_WITH_USERBUFFERS)
    message(STATUS "userbuffers support enabled")
    add_subdirectory(pytorch/csrc/userbuffers)
endif()
