Add build doc

CMakeLists.txt

@@ -5,6 +5,11 @@ include(cmake/Utils.cmake)
 set_default_configuration_release()
 msvc_use_static_runtime()
 
+# Options
+option(USE_CUDA  "Build with CUDA" OFF)
+option(USE_CUDNN "Build with CUDNN" OFF)
+option(USE_TENSORRT "Build with Tensor RT" OFF)
+
 # Use RPATH on Mac OS X as flexible mechanism for locating dependencies
 # See https://blog.kitware.com/upcoming-in-cmake-2-8-12-osx-rpath-support/
 set(CMAKE_MACOSX_RPATH TRUE)
@@ -56,11 +61,7 @@ FILE(GLOB_RECURSE DLR_INC
     "${TREELITE_SRC}/include/*.h"
 )
 
-set(USE_CUDA OFF)
-set(USE_CUDNN OFF)
-set(USE_TENSORRT OFF)
-
-if(USE_CUDA STREQUAL "ON")
+if(USE_CUDA)
     message("USING CUDA")
     set(USE_CUDA "/usr/local/cuda-9.0")
     set(CUDA_TOOLKIT_ROOT_DIR ${USE_CUDA})
@@ -84,7 +85,7 @@ if(USE_CUDA STREQUAL "ON")
     file(GLOB RUNTIME_CUDA_SRCS ${TVM_SRC}/src/runtime/cuda/*.cc)
     list(APPEND DLR_SRC ${RUNTIME_CUDA_SRCS})
 endif()
-if(USE_CUDNN STREQUAL "ON")
+if(USE_CUDNN)
     message("USING CUDNN")
     set(USE_CUDNN ${USE_CUDA})
     set(CUDNN_TOOLKIT_ROOT_DIR ${USE_CUDNN})
@@ -100,7 +101,7 @@ if(USE_CUDNN STREQUAL "ON")
     file(GLOB CONTRIB_CUDNN_SRCS ${TVM_SRC}/src/contrib/cudnn/*.cc)
     list(APPEND RUNTIME_SRCS ${CONTRIB_CUDNN_SRCS})
 endif()
-if(USE_TENSORRT STREQUAL "ON")
+if(USE_TENSORRT)
     message("USING TENSORRT")
     set(USE_TENSORRT "/home/ubuntu/TensorRT-4.0.1.6")
     set(TENSORRT_ROOT_DIR ${USE_TENSORRT})

README.md

@@ -3,6 +3,7 @@
 DLR is a compact, common runtime for deep learning models and decision tree models compiled by [AWS SageMaker Neo](https://aws.amazon.com/sagemaker/neo/), [TVM](https://tvm.ai/), or [Treelite](https://treelite.readthedocs.io/en/latest/install.html). DLR uses the TVM runtime, Treelite runtime, NVIDIA TensorRT™, and can include other hardware-specific runtimes. DLR provides unified Python/C++ APIs for loading and running compiled models on various devices. DLR currently supports platforms from Intel, NVIDIA, and ARM, with support for Xilinx, Cadence, and Qualcomm coming soon.
 
 ## Documentation
+For instructions on installilng DLR, please refer to [Installing DLR](https://neo-ai-dlr.readthedocs.io/en/latest/install.html)
 
 For instructions on using DLR, please refer to [Amazon SageMaker Neo – Train Your Machine Learning Models Once, Run Them Anywhere](https://aws.amazon.com/blogs/aws/amazon-sagemaker-neo-train-your-machine-learning-models-once-run-them-anywhere/)
 

doc/Doxyfile

@@ -7,7 +7,7 @@ CASE_SENSE_NAMES = NO
 ENABLE_PREPROCESSING   = YES
 MACRO_EXPANSION        = YES
 EXPAND_ONLY_PREDEF     = YES
-PREDEFINED             = 
+PREDEFINED             = DLR_DLL=
 INPUT            = ../include/ ../src/
 EXAMPLE_PATH     = ../
 RECURSIVE        = YES

doc/index.rst

@@ -10,6 +10,7 @@ Contents
   :maxdepth: 2
   :titlesonly:
 
+  install
   python-api
   c-api
   Internal docs <http://neo-ai-dlr.readthedocs.io/en/latest/dev/>

doc/install.rst

@@ -0,0 +1,116 @@
+##############
+Installing DLR
+##############
+
+.. contents:: Contents
+  :local:
+  :backlinks: none
+
+************************
+Building DLR from source
+************************
+
+Building DLR consists of two steps:
+
+1. Build the shared library from C++ code (``libdlr.so`` for Linux, ``libdlr.dylib`` for Mac OSX, and ``dlr.dll`` for Windows).
+2. Then install the Python package ``dlr``.
+
+.. note:: Use of Git submodules
+
+  DLR uses Git submodules to manage dependencies. So when you clone the repo, remember to specify ``--recursive`` option:
+
+  .. code-block:: bash
+
+    git clone --recursive https://github.com/neo-ai/neo-ai-dlr
+
+Building on Linux
+=================
+
+Ensure that all necessary software packages are installed: GCC (or Clang), CMake, and Python. For example, in Ubuntu, you can run
+
+.. code-block:: bash
+
+  sudo apt-get update
+  sudo apt-get install -y python3 python3-pip gcc build-essential cmake
+
+To build, create a subdirectory ``build`` and invoke CMake:
+
+.. code-block:: bash
+
+  mkdir build
+  cd build
+  cmake ..
+
+Once CMake is done generating a Makefile, run GNU Make to compile:
+
+.. code-block:: bash
+
+  make -j4         # Use 4 cores to compile sources in parallel
+
+By default, DLR will be built with CPU support only. To enable support for NVIDIA GPUs, enable CUDA, CUDNN, and TensorRT by calling CMake with extra options:
+
+.. code-block:: bash
+
+  cmake .. -DUSE_CUDA=ON -DUSE_TENSORRT=ON -DUSE_CUDNN=ON
+  make -j4
+
+You will need to install NVIDIA CUDA toolkits and drivers beforehand.
+
+Once the compilation is completed, install the Python package by running ``setup.py``:
+
+.. code-block:: bash
+
+  cd python
+  python3 setup.py install
+
+Building on Mac OS X
+====================
+
+Install GCC and CMake from `Homebrew <https://brew.sh/>`_:
+
+.. code-block:: bash
+
+  brew update
+  brew install cmake gcc@8
+
+To ensure that Homebrew GCC is used (instead of default Apple compiler), specify environment variables ``CC`` and ``CXX`` when invoking CMake:
+
+.. code-block:: bash
+
+  mkbir build
+  cd build
+  CC=gcc-8 CXX=g++-8 cmake ..
+  make -j4
+
+NVIDIA GPUs are not supported for Mac OS X target.
+
+Once the compilation is completed, install the Python package by running ``setup.py``:
+
+.. code-block:: bash
+
+  cd python
+  python3 setup.py install
+
+Building on Windows
+===================
+
+DLR requires `Visual Studio 2017 <https://visualstudio.microsoft.com/downloads/>`_ as well as `CMake <https://cmake.org/>`_.
+
+In the DLR directory, first run CMake to generate a Visual Studio project:
+
+.. code-block:: cmd
+
+  mkdir build
+  cd build
+  cmake .. -G"Visual Studio 15 2017 Win64"
+
+If CMake run was successful, you should be able to find the solution file ``dlr.sln``. Open it with Visual Studio. To build, choose **Build Solution** on the **Build** menu.
+
+NVIDIA GPU is not yet supported for Windows target.
+
+Once the compilation is completed, install the Python package by running ``setup.py``:
+
+.. code-block:: bash
+
+  cd python
+  python3 setup.py install