I installed cuda 10.0 on my machine, and I added the lib64 to LD_LIBRARY_PATH.
I installed tensorrt 5.1.5 following the offical steps https://docs.nvidia.com/deeplearning/sdk/tensorrt-install-guide/index.html#installing-rpm,,but it reported errors after I did “yum install tensorrt”:
yum install tensorrt
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
- base: mirrors.aliyun.com
- extras: mirrors.aliyun.com
- updates: mirrors.aliyun.com
Resolving Dependencies
→ Running transaction check
—> Package tensorrt.x86_64 0:5.1.5.0-1.cuda10.0 will be installed
→ Processing Dependency: libnvinfer5 = 5.1.5 for package: tensorrt-5.1.5.0-1.cuda10.0.x86_64
→ Processing Dependency: libnvinfer-samples = 5.1.5 for package: tensorrt-5.1.5.0-1.cuda10.0.x86_64
→ Processing Dependency: libnvinfer-devel = 5.1.5 for package: tensorrt-5.1.5.0-1.cuda10.0.x86_64
→ Running transaction check
—> Package libnvinfer-devel.x86_64 0:5.1.5-1.cuda10.0 will be installed
→ Processing Dependency: cuda-cudart-dev-10-0 for package: libnvinfer-devel-5.1.5-1.cuda10.0.x86_64
→ Processing Dependency: cuda-cublas-dev-10-0 for package: libnvinfer-devel-5.1.5-1.cuda10.0.x86_64
—> Package libnvinfer-samples.x86_64 0:5.1.5-1.cuda10.0 will be installed
—> Package libnvinfer5.x86_64 0:5.1.5-1.cuda10.0 will be installed
→ Processing Dependency: cuda-cudart-10-0 for package: libnvinfer5-5.1.5-1.cuda10.0.x86_64
→ Processing Dependency: cuda-cublas-10-0 for package: libnvinfer5-5.1.5-1.cuda10.0.x86_64
→ Finished Dependency Resolution
Error: Package: libnvinfer-devel-5.1.5-1.cuda10.0.x86_64 (nv-tensorrt-cuda10.0-trt5.1.5.0-ga-20190427)
Requires: cuda-cudart-dev-10-0
Error: Package: libnvinfer5-5.1.5-1.cuda10.0.x86_64 (nv-tensorrt-cuda10.0-trt5.1.5.0-ga-20190427)
Requires: cuda-cublas-10-0
Error: Package: libnvinfer5-5.1.5-1.cuda10.0.x86_64 (nv-tensorrt-cuda10.0-trt5.1.5.0-ga-20190427)
Requires: cuda-cudart-10-0
Error: Package: libnvinfer-devel-5.1.5-1.cuda10.0.x86_64 (nv-tensorrt-cuda10.0-trt5.1.5.0-ga-20190427)
Requires: cuda-cublas-dev-10-0
You could try using --skip-broken to work around the problem
Could anyone help me ?