This article mainly records the environment configuration process of TensorRT8.6!
Official Documentation: NVIDIA TensorRT - NVIDIA Docs
Documentation for TensorRT related versions: Documentation Archives :: NVIDIA Deep Learning TensorRT Documentation
1. Download CUDA and cudann
CUDA下载:CUDA Toolkit Archive | NVIDIA Developer
CUDA installation: ( I am using CUDA 11.0)
Baidu Netdisk (CUDA 11.0):
Link: https://pan.baidu.com/s/1ZpPkNRDtcbQURIEgpF7t5Q
Extraction code: dn6q
1. (Windows version) installation
① CUDA installation and testing
Double-click the downloaded exe file
Just install it by default all the way, and finally test it:
nvcc -V
In this way, CUDA 11.0 is installed!
②cudann installation and testing
cudann download: cuDNN Archive | NVIDIA Developer
Baidu network disk (cudann8.0.2):
Link: https://pan.baidu.com/s/13JDfexry0hP1GV0fgnbbBg
Extraction code: r83z
After the download is complete, it is a compressed file. After decompression, copy (bin, include, lib) to the cuda directory where the installation is successful.
My path is this C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0
have a test:
In this way, cudann is installed successfully!
2. Download and installation of TensorRT
Download link: TensorRT SDK | NVIDIA Developer
Baidu Netdisk (TensorRT8.6):
Link: https://pan.baidu.com/s/1KFkUFNZhNfj0Wo0fKSLbNg
Extraction code: tec5
1. Download TensorRT
1. Log in to your account first and then click Download
2. Select TensorRT 8
2. Install TensorRT
Decompress the downloaded TensorRT compressed package (I decompressed it to the D drive), so that the installation is actually completed, it is as simple as that!
Next configure the environment variables:
D:\TensorRT-8.6.0.12\lib
3. Configure the VS2017 development environment
Note: It needs to be set to Release x64!
1. Contains directory configuration
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0\include
D:\TensorRT-8.6.0.12\include
2. Library directory
D:\TensorRT-8.6.0.12\lib
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0\lib\x64
3. Linker, additional dependencies
nvinfer.lib
nvinfer_dispatch.lib
nvinfer_lean.lib
nvinfer_plugin.lib
nvinfer_vc_plugin.lib
nvonnxparser.lib
nvparsers.lib
cublas.lib
cublasLt.lib
cuda.lib
cudadevrt.lib
cudart.lib
cudart_static.lib
cudnn.lib
cudnn_adv_infer.lib
cudnn_adv_infer64_8.lib
cudnn_adv_train.lib
cudnn_adv_train64_8.lib
cudnn_cnn_infer.lib
cudnn_cnn_infer64_8.lib
cudnn_cnn_train.lib
cudnn_cnn_train64_8.lib
cudnn_ops_infer.lib
cudnn_ops_infer64_8.lib
cudnn_ops_train.lib
cudnn_ops_train64_8.lib
cudnn64_8.lib
cufft.lib
cufftw.lib
curand.lib
cusolver.lib
cusolverMg.lib
cusparse.lib
nppc.lib
nppial.lib
nppicc.lib
nppidei.lib
nppif.lib
nppig.lib
nppim.lib
nppist.lib
nppisu.lib
nppitc.lib
npps.lib
nvblas.lib
nvjpeg.lib
nvml.lib
nvrtc.lib
OpenCL.lib
Test code:
#include <iostream>
#include "NvInfer.h"
#include "NvOnnxParser.h"
using namespace nvinfer1;
using namespace nvonnxparser;
class Logger : public ILogger
{
void log(Severity severity, const char* msg) noexcept
{
if (severity != Severity::kINFO)
std::cout << msg << std::endl;
}
}gLogger;
int main(int argc, char** argv)
{
auto builder = createInferBuilder(gLogger);
builder->getLogger()->log(nvinfer1::ILogger::Severity::kERROR, "Create Builder ...");
return 0;
}
In this way, the Windows environment is built and configured!