/TensorRT

TensorRT is a C++ library for high performance inference on NVIDIA GPUs and deep learning accelerators.

Primary LanguageC++Apache License 2.0Apache-2.0

License Documentation

TensorRT Open Source Software

This repository contains the Open Source Software (OSS) components of NVIDIA TensorRT. Included are the sources for TensorRT plugins and parsers (Caffe and ONNX), as well as sample applications demonstrating usage and capabilities of the TensorRT platform. These open source software components are a subset of the TensorRT General Availability (GA) release with some extensions and bug-fixes.

Build

Prerequisites

To build the TensorRT-OSS components, you will first need the following software packages.

TensorRT GA build

System Packages

Optional Packages

Downloading TensorRT Build

  1. Download TensorRT OSS

    On Linux: Bash

    git clone -b master https://github.com/nvidia/TensorRT TensorRT
    cd TensorRT
    git submodule update --init --recursive

    On Windows: Powershell

    git clone -b master https://github.com/nvidia/TensorRT TensorRT
    cd TensorRT
    git submodule update --init --recursive
  2. Specify the TensorRT Release build

    If using NVIDIA build containers, TensorRT is preinstalled under /usr/lib/x86_64-linux-gnu.

    Else download and extract the TensorRT build from NVIDIA Developer Zone.

    Example: Ubuntu 18.04 on x86-64 with cuda-11.1

    cd ~/Downloads
    tar -xvzf TensorRT-7.2.3.4.Ubuntu-18.04.x86_64-gnu.cuda-11.1.cudnn8.1.tar.gz
    export TRT_LIBPATH=`pwd`/TensorRT-7.2.3.4

    Example: Windows on x86-64 with cuda-11.0

    cd ~\Downloads
    Expand-Archive .\TensorRT-7.2.3.4.Windows10.x86_64.cuda-11.0.cudnn8.1.zip
    $Env:TRT_LIBPATH = '$(Get-Location)\TensorRT-7.2.3.4'
    $Env:PATH += 'C:\Program Files (x86)\Microsoft Visual Studio\2017\Professional\MSBuild\15.0\Bin\'
  3. (Optional) JetPack SDK for Jetson builds

    Using the JetPack SDK manager, download the host components. Steps:

    1. Download and launch the SDK manager. Login with your developer account.
    2. Select the platform and target OS (example: Jetson AGX Xavier, Linux Jetpack 4.4), and click Continue.
    3. Under Download & Install Options change the download folder and select Download now, Install later. Agree to the license terms and click Continue.
    4. Move the extracted files into the <TensorRT-OSS>/docker/jetpack_files folder.

Setting Up The Build Environment

For native builds, install the prerequisite System Packages. Alternatively (recommended for non-Windows builds), install Docker and generate a build container as described below:

  1. Generate the TensorRT-OSS build container.

    The TensorRT-OSS build container can be generated using the Dockerfiles and build script included with TensorRT-OSS. The build container is bundled with packages and environment required for building TensorRT OSS.

    Example: Ubuntu 18.04 on x86-64 with cuda-11.1

    ./docker/build.sh --file docker/ubuntu-18.04.Dockerfile --tag tensorrt-ubuntu-1804 --cuda 11.1

    Example: Ubuntu 18.04 cross-compile for PowerPC with cuda-11.0

    ./docker/build.sh --file docker/ubuntu-cross-ppc64le.Dockerfile --tag tensorrt-ubuntu-ppc --cuda 11.0

    Example: CentOS/RedHat 7 on x86-64 with cuda-11.0

    ./docker/build.sh --file docker/centos-7.Dockerfile --tag tensorrt-centos --cuda 11.0

    Example: Ubuntu 18.04 cross-compile for Jetson (arm64) with cuda-10.2 (JetPack SDK)

    ./docker/build.sh --file docker/ubuntu-cross-aarch64.Dockerfile --tag tensorrt-cross-jetpack --cuda 10.2
  2. Launch the TensorRT-OSS build container.

    Example: Ubuntu 18.04 build container

    ./docker/launch.sh --tag tensorrt-ubuntu-1804 --gpus all

    NOTE:

    1. Use the tag corresponding to the build container.
    2. To run TensorRT/CUDA programs in the build container, install NVIDIA Container Toolkit. Docker versions < 19.03 require nvidia-docker2 and --runtime=nvidia flag for docker run commands. On versions >= 19.03, you need the nvidia-container-toolkit package and --gpus all flag.
    3. sudo password for Ubuntu containers is 'nvidia'.

Building TensorRT-OSS

  • Generate Makefiles or VS project (Windows) and build.

    Example: Linux (x86-64) build with default cuda-11.1

     cd $TRT_OSSPATH
     mkdir -p build && cd build
     cmake .. -DTRT_LIB_DIR=$TRT_LIBPATH -DTRT_OUT_DIR=`pwd`/out
     make -j$(nproc)

    Example: Native build on Jetson (arm64) with cuda-10.2

    cd $TRT_OSSPATH
    mkdir -p build && cd build
    cmake .. -DTRT_LIB_DIR=$TRT_LIBPATH -DTRT_OUT_DIR=`pwd`/out -DTRT_PLATFORM_ID=aarch64 -DCUDA_VERSION=10.2
    make -j$(nproc)

    Example: Ubuntu 18.04 Cross-Compile for Jetson (arm64) with cuda-10.2 (JetPack)

     cd $TRT_OSSPATH
     mkdir -p build && cd build
     cmake .. -DTRT_LIB_DIR=$TRT_LIBPATH -DTRT_OUT_DIR=`pwd`/out -DCMAKE_TOOLCHAIN_FILE=$TRT_OSSPATH/cmake/toolchains/cmake_aarch64.toolchain -DCUDA_VERSION=10.2
     make -j$(nproc)

    Example: Cross-Compile for QNX with cuda-10.2

     cd $TRT_OSSPATH
     mkdir -p build && cd build
     cmake .. -DTRT_LIB_DIR=$TRT_LIBPATH -DTRT_OUT_DIR=`pwd`/out -DCMAKE_TOOLCHAIN_FILE=$TRT_OSSPATH/cmake/toolchains/cmake_qnx.toolchain -DCUDA_VERSION=10.2
     make -j$(nproc)

    Example: Windows (x86-64) build in Powershell

     cd $Env:TRT_OSSPATH
     mkdir -p build ; cd build
     cmake .. -DTRT_LIB_DIR=$Env:TRT_LIBPATH -DTRT_OUT_DIR='$(Get-Location)\out' -DCMAKE_TOOLCHAIN_FILE=..\cmake\toolchains\cmake_x64_win.toolchain
     msbuild ALL_BUILD.vcxproj

    NOTE:

    1. The default CUDA version used by CMake is 11.1. To override this, for example to 10.2, append -DCUDA_VERSION=10.2 to the cmake command.
    2. If samples fail to link on CentOS7, create this symbolic link: ln -s $TRT_OUT_DIR/libnvinfer_plugin.so $TRT_OUT_DIR/libnvinfer_plugin.so.7
  • Required CMake build arguments are:

    • TRT_LIB_DIR: Path to the TensorRT installation directory containing libraries.
    • TRT_OUT_DIR: Output directory where generated build artifacts will be copied.
  • Optional CMake build arguments:

    • CMAKE_BUILD_TYPE: Specify if binaries generated are for release or debug (contain debug symbols). Values consists of [Release] | Debug
    • CUDA_VERISON: The version of CUDA to target, for example [11.1].
    • CUDNN_VERSION: The version of cuDNN to target, for example [8.0].
    • NVCR_SUFFIX: Optional nvcr/cuda image suffix. Set to "-rc" for CUDA11 RC builds until general availability. Blank by default.
    • PROTOBUF_VERSION: The version of Protobuf to use, for example [3.0.0]. Note: Changing this will not configure CMake to use a system version of Protobuf, it will configure CMake to download and try building that version.
    • CMAKE_TOOLCHAIN_FILE: The path to a toolchain file for cross compilation.
    • BUILD_PARSERS: Specify if the parsers should be built, for example [ON] | OFF. If turned OFF, CMake will try to find precompiled versions of the parser libraries to use in compiling samples. First in ${TRT_LIB_DIR}, then on the system. If the build type is Debug, then it will prefer debug builds of the libraries before release versions if available.
    • BUILD_PLUGINS: Specify if the plugins should be built, for example [ON] | OFF. If turned OFF, CMake will try to find a precompiled version of the plugin library to use in compiling samples. First in ${TRT_LIB_DIR}, then on the system. If the build type is Debug, then it will prefer debug builds of the libraries before release versions if available.
    • BUILD_SAMPLES: Specify if the samples should be built, for example [ON] | OFF.
    • CUB_VERSION: The version of CUB to use, for example [1.8.0].
    • GPU_ARCHS: GPU (SM) architectures to target. By default we generate CUDA code for all major SMs. Specific SM versions can be specified here as a quoted space-separated list to reduce compilation time and binary size. Table of compute capabilities of NVIDIA GPUs can be found here. Examples:
      • NVidia A100: -DGPU_ARCHS="80"
      • Tesla T4, GeForce RTX 2080: -DGPU_ARCHS="75"
      • Titan V, Tesla V100: -DGPU_ARCHS="70"
      • Multiple SMs: -DGPU_ARCHS="80 75"
    • TRT_PLATFORM_ID: Bare-metal build (unlike containerized cross-compilation) on non Linux/x86 platforms must explicitly specify the target platform. Currently supported options: x86_64 (default), aarch64

References

TensorRT Resources

Known Issues

TensorRT 7.2.3.4

  • None