Closed Shaked closed 4 years ago
@Shaked I tried to the following :
sudo cmake .. -DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/cmake_aarch64.toolchain -DCUBLASLT_LIB=/usr/local/cuda/targets/aarch64-linux/lib/libcublas.so -DBUILD_SAMPLES=OFF -DBUILD_PLUGINS=OFF -DTRT_LIB_DIR=$TRT_RELEASE/lib -DCUDA_VERSION=10.2 -DCMAKE_CUDA_COMPILER=/usr/local/cuda-10.2/bin/nvcc
It did better and protopuf got it the 100%. But it fails to link CUDA libraries. So now my guess is that I'm still missing cmake
varibale to tell the cross linker where are the libs?
[ 97%] Linking CXX executable onnx2trt
/usr/lib/gcc-cross/aarch64-linux-gnu/7/../../../../aarch64-linux-gnu/bin/ld: warning: libcudnn.so.8, needed by /pdk_files/tensorrt/lib/libnvinfer.so, not found (try using -rpath or -rpath-link)
/usr/lib/gcc-cross/aarch64-linux-gnu/7/../../../../aarch64-linux-gnu/bin/ld: warning: libcudart.so.10.2, needed by /pdk_files/tensorrt/lib/libnvinfer.so, not found (try using -rpath or -rpath-link)
/usr/lib/gcc-cross/aarch64-linux-gnu/7/../../../../aarch64-linux-gnu/bin/ld: warning: libnvrtc.so.10.2, needed by /pdk_files/tensorrt/lib/libnvinfer.so, not found (try using -rpath or -rpath-link)
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaGetDevice'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaDeviceSynchronize'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaPeekAtLastError'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcCreateProgram@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDestroy@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDestroyPoolingDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreateTensorDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `__cudaRegisterFunction@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetActivationDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `__cudaUnregisterFatBinary@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaFreeHost'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDestroyOpTensorDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaMemcpy2DAsync@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDestroyConvolutionDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreateOpTensorDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `nvrtcAddNameExpression@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetConvolutionForwardAlgorithmMaxCount@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `__cudaPushCallConfiguration'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaGetErrorName'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetConvolutionForwardAlgorithm_v7@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaEventDestroy'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaCreateTextureObject@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetBatchNormalizationForwardTrainingExWorkspaceSize@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaDeviceGetAttribute'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaDestroyTextureObject@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDestroyTensorDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaHostAlloc'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaEventCreateWithFlags@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `__cudaUnregisterFatBinary'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaFuncGetAttributes'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetDropoutDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnBatchNormalizationBackwardEx@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreateActivationDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaPeekAtLastError@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaEventCreateWithFlags'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetTensorNdDescriptorEx@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `__cudaRegisterFatBinary'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnPoolingBackward@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnGetRNNWorkspaceSize@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnGetFilterNdDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetConvolution2dDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetConvolutionGroupCount@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcDestroyProgram@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnConvolutionBackwardData@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcGetPTXSize@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaGetErrorString@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnGetRNNLinLayerMatrixParams@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudnnCreateTensorDescriptor'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetTensorNdDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetConvolutionMathType@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetTensor4dDescriptorEx@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `__cudaPopCallConfiguration@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `__cudaRegisterVar@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaRuntimeGetVersion@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaGetDeviceProperties@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaGetLastError'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaStreamCreateWithPriority'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetConvolutionBackwardDataAlgorithmMaxCount@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSoftmaxForward@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaMalloc'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaMemcpy'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `__cudaRegisterFatBinary@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `__cudaRegisterVar'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreate@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudnnDestroyTensorDescriptor'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaFree@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcCompileProgram@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaMemcpyAsync'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnGetRNNLinLayerBiasParams@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaStreamDestroy@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreateRNNDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaGetDevice@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnGetErrorString@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaStreamWaitEvent@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaEventDestroy@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaEventSynchronize@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaMallocHost'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaStreamWaitEvent'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnGetRNNParamsSize@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaStreamCreate'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaGetDeviceProperties'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDestroyActivationDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDestroyFilterDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetStream@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaDeviceGetAttribute@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaGetDeviceCount@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcGetProgramLog@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetRNNMatrixMathType@libcudnn.so.8'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetProperty@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaStreamCreate@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaMemset'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaEventElapsedTime@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaEventRecord@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaDeviceSynchronize@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnPoolingForward@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `__cudaRegisterFatBinaryEnd'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `__cudaPopCallConfiguration'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnConvolutionBackwardFilter@libcudnn.so.8'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetBatchNormalizationBackwardExWorkspaceSize@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `__cudaRegisterFatBinaryEnd@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetConvolutionMathType@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnGetVersion@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaStreamSynchronize'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `nvrtcGetLoweredName@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDestroyRNNDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaLaunchKernel@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreateDropoutDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaMemcpyToSymbol@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetFilter4dDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnGetConvolutionBackwardDataWorkspaceSize@libcudnn.so.8'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnOpTensor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetPooling2dDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaStreamAddCallback@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetLRNDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnAddTensor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDestroyLRNDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetBatchNormalizationTrainingExReserveSpaceSize@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDropoutGetStatesSize@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetConvolutionNdDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnGetConvolutionForwardWorkspaceSize@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcVersion@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudnnSetStream'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnRNNForwardInference@libcudnn.so.8'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnSetPoolingNdDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnLRNCrossChannelForward@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetFilterNdDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreateFilterDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnActivationForward@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaLaunchKernel'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetRNNDescriptor_v6@libcudnn.so.8'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetConvolutionBackwardFilterAlgorithmMaxCount@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaEventElapsedTime'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaFuncSetAttribute@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaGetErrorString'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaMemsetAsync@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetTensorNdDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaStreamDestroy'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudnnSetTensor4dDescriptor'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetTensor4dDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaEventSynchronize'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnConvolutionForward@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaMemsetAsync'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaMemGetInfo@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnGetTensor4dDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaMemcpy@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaEventCreate'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaDeviceReset'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnDeriveBNTensorDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcGetErrorString@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudnnBatchNormalizationForwardTraining'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `__cudaRegisterFunction'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaMalloc@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcGetPTX@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnBatchNormalizationForwardTrainingEx@libcudnn.so.8'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetConvolutionBackwardFilterAlgorithm_v7@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnDestroyDropoutDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreateLRNDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaStreamSynchronize@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetConvolutionBackwardDataAlgorithm_v7@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcGetProgramLogSize@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaFree'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaEventRecord'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaFuncSetAttribute'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreatePoolingDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaCreateChannelDesc@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `__cudaPushCallConfiguration@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreateConvolutionDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaMemcpyAsync@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnSetOpTensorDescriptor@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaGetLastError@libcudart.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnTransformTensor@libcudnn.so.8'
collect2: error: ld returned 1 exit status
parsers/onnx/CMakeFiles/onnx2trt.dir/build.make:90: recipe for target 'parsers/onnx/onnx2trt' failed
make[2]: *** [parsers/onnx/onnx2trt] Error 1
CMakeFiles/Makefile2:535: recipe for target 'parsers/onnx/CMakeFiles/onnx2trt.dir/all' failed
make[1]: *** [parsers/onnx/CMakeFiles/onnx2trt.dir/all] Error 2
Makefile:151: recipe for target 'all' failed
make: *** [all] Error 2
@oak-tree
Your solution seems good!
I used the following:
sudo cmake .. \
-DCMAKE_TOOLCHAIN_FILE="../cmake/toolchains/cmake_aarch64.toolchain" \
-DCUBLASLT_LIB="/usr/local/cuda/targets/aarch64-linux/lib/libcublas.so" \
-DBUILD_SAMPLES="OFF" \
-DBUILD_PLUGINS="OFF" \
-DTRT_LIB_DIR="/pdk_files/tensorrt/lib" \
-DCUDA_VERSION="10.2" \
-DCMAKE_CUDA_COMPILER="/usr/local/cuda-10.2/bin/nvcc" \
-DCUDNN_VERSION="8.0" \
-DPROTOBUF_VERSION="3.8.0" \
-DTRT_OUT_DIR="`pwd`/out" \
-DCUDNN_LIB="/usr/local/cuda/lib64"
Then I ran
sudo CC=aarch64-linux-gnu-gcc make -j8 &> x.log
At some point I see things like:
[ 7%] Built target third_party.protobuf
[ 16%] Built target third_party.protobuf_aarch64
[ 19%] Built target gen_onnx_proto
[ 19%] Built target caffe_proto
[ 25%] Built target onnx_proto
[ 32%] Built target nvonnxparser_static
[ 40%] Built target nvonnxparser
[ 68%] Built target nvcaffeparser_static
[ 95%] Built target nvcaffeparser
[ 97%] Linking CXX executable onnx2trt
[ 97%] Building CXX object parsers/onnx/CMakeFiles/getSupportedAPITest.dir/getSupportedAPITest.cpp.o
[ 98%] Building CXX object parsers/onnx/CMakeFiles/getSupportedAPITest.dir/ModelImporter.cpp.o
In file included from /workspace/TensorRT/build/parsers/onnx/third_party/onnx/onnx/onnx-ml.pb.h:2:0,
from /workspace/TensorRT/parsers/onnx/third_party/onnx/onnx/onnx_pb.h:50,
from /workspace/TensorRT/parsers/onnx/ShapedWeights.hpp:26,
from /workspace/TensorRT/parsers/onnx/onnx2trt.hpp:26,
from /workspace/TensorRT/parsers/onnx/ImporterContext.hpp:25,
from /workspace/TensorRT/parsers/onnx/ModelImporter.hpp:25,
from /workspace/TensorRT/parsers/onnx/ModelImporter.cpp:23:
/workspace/TensorRT/build/parsers/onnx/third_party/onnx/onnx/onnx_onnx2trt_onnx-ml.pb.h:17:2: error: #error This file was generated by an older version of protoc which is
#error This file was generated by an older version of protoc which is
^~~~~
/workspace/TensorRT/build/parsers/onnx/third_party/onnx/onnx/onnx_onnx2trt_onnx-ml.pb.h:18:2: error: #error incompatible with your Protocol Buffer headers. Please
#error incompatible with your Protocol Buffer headers. Please
^~~~~
/workspace/TensorRT/build/parsers/onnx/third_party/onnx/onnx/onnx_onnx2trt_onnx-ml.pb.h:19:2: error: #error regenerate this file with a newer version of protoc.
#error regenerate this file with a newer version of protoc.
^~~~~
In file included from /workspace/TensorRT/build/parsers/onnx/third_party/onnx/onnx/onnx-ml.pb.h:2:0,
from /workspace/TensorRT/parsers/onnx/third_party/onnx/onnx/onnx_pb.h:50,
from /workspace/TensorRT/parsers/onnx/onnx_utils.hpp:28,
from /workspace/TensorRT/parsers/onnx/getSupportedAPITest.cpp:29:
/workspace/TensorRT/build/parsers/onnx/third_party/onnx/onnx/onnx_onnx2trt_onnx-ml.pb.h:17:2: error: #error This file was generated by an older version of protoc which is
#error This file was generated by an older version of protoc which is
^~~~~
/workspace/TensorRT/build/parsers/onnx/third_party/onnx/onnx/onnx_onnx2trt_onnx-ml.pb.h:18:2: error: #error incompatible with your Protocol Buffer headers. Please
#error incompatible with your Protocol Buffer headers. Please
^~~~~
/workspace/TensorRT/build/parsers/onnx/third_party/onnx/onnx/onnx_onnx2trt_onnx-ml.pb.h:19:2: error: #error regenerate this file with a newer version of protoc.
#error regenerate this file with a newer version of protoc.
^~~~~
/usr/lib/gcc-cross/aarch64-linux-gnu/7/../../../../aarch64-linux-gnu/bin/ld: warning: libcudnn.so.8, needed by /pdk_files/tensorrt/lib/libnvinfer.so, not found (try using -rpath or -rpath-link)
/usr/lib/gcc-cross/aarch64-linux-gnu/7/../../../../aarch64-linux-gnu/bin/ld: warning: libcudart.so.10.2, needed by /pdk_files/tensorrt/lib/libnvinfer.so, not found (try using -rpath or -rpath-link)
/usr/lib/gcc-cross/aarch64-linux-gnu/7/../../../../aarch64-linux-gnu/bin/ld: warning: libnvrtc.so.10.2, needed by /pdk_files/tensorrt/lib/libnvinfer.so, not found (try using -rpath or -rpath-link)
third_party/onnx/libonnx_proto.a(onnx_onnx2trt_onnx-ml.pb.cc.o): In function `onnx2trt_onnx::OperatorSetIdProto::ByteSize() const':
onnx_onnx2trt_onnx-ml.pb.cc:(.text+0x21c): undefined reference to `google_private::protobuf::io::CodedOutputStream::VarintSize64(unsigned long)'
Attached full log:
I have just noticed that cmake
does contain a problem although it finishes successfully:
sudo cmake .. \
> -DCMAKE_TOOLCHAIN_FILE="../cmake/toolchains/cmake_aarch64.toolchain" \
> -DCUBLASLT_LIB="/usr/local/cuda/targets/aarch64-linux/lib/libcublas.so" \
> -DBUILD_SAMPLES="OFF" \
> -DBUILD_PLUGINS="OFF" \
> -DTRT_LIB_DIR="/pdk_files/tensorrt/lib" \
> -DCUDA_VERSION="10.2" \
> -DCMAKE_CUDA_COMPILER="/usr/local/cuda-10.2/bin/nvcc" \
> -DCUDNN_VERSION="8.0" \
> -DPROTOBUF_VERSION="3.8.0" \
> -DTRT_OUT_DIR="`pwd`/out" \
> -DTENSORRT_LIBRARY_INFER_PLUGIN="/pdk_files/tensorrt/lib/stubs" \
> -DCUDNN_LIB="/usr/local/cuda/lib64"
Building for TensorRT version: 7.1.3, library version: 7
-- The CXX compiler identification is GNU 7.5.0
-- The CUDA compiler identification is NVIDIA 10.2.89
-- Targeting TRT Platform: aarch64
-- CUDA version set to 10.2
-- cuDNN version set to 8.0
-- Protobuf version set to 3.8.0
-- Looking for C++ include pthread.h
-- Looking for C++ include pthread.h - found
-- Looking for pthread_create
-- Looking for pthread_create - not found
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE
-- Found PkgConfig: /usr/bin/pkg-config (found version "0.29.1")
-- Checking for one of the modules 'zlib'
-- Setting up another Protobuf build for cross compilation targeting aarch64-Linux
-- Using libprotobuf /workspace/TensorRT/build_shaked/third_party.protobuf_aarch64/lib/libprotobuf.a
-- ========================= Importing and creating target nvinfer ==========================
-- Looking for library nvinfer
-- Library that was found /pdk_files/tensorrt/lib/libnvinfer.so
-- ==========================================================================================
-- ========================= Importing and creating target nvuffparser ==========================
-- Looking for library nvparsers
-- Library that was found /pdk_files/tensorrt/lib/libnvparsers.so
-- ==========================================================================================
CMake Warning at CMakeLists.txt:159 (message):
Detected CUDA version is < 11.0. SM80 not supported.
-- GPU_ARCHS is not defined. Generating CUDA code for default SMs: 35;53;61;70;75
-- ========================= Importing and creating target nvinfer_plugin ==========================
-- Looking for library nvinfer_plugin
-- Library that was found nvinfer_plugin_LIB_PATH-NOTFOUND
-- ==========================================================================================
-- Protobuf proto/trtcaffe.proto -> proto/trtcaffe.pb.cc proto/trtcaffe.pb.h
-- /workspace/TensorRT/build_shaked/parsers/caffe
-- The C compiler identification is GNU 7.5.0
-- Build type not set - defaulting to Release
-- Looking for library nvinfer_plugin -- Library that was found nvinfer_plugin_LIB_PATH-NOTFOUND --
Means it doesn't find it for some reason.
I have tried to clone this repo again and use the release/7.1
tag.
After doing that I ran the following cmake:
sudo cmake .. \
-DCMAKE_TOOLCHAIN_FILE="/workspace/TensorRT/SHAKED/TensorRT/cmake/toolchains/cmake_aarch64.toolchain" \
-DCUBLASLT_LIB="/usr/local/cuda/targets/aarch64-linux/lib/libcublas.so" \
-DBUILD_SAMPLES="OFF" \
-DBUILD_PLUGINS="OFF" \
-DTRT_LIB_DIR="/pdk_files/tensorrt/lib" \
-DCUDA_VERSION="10.2" \
-DCMAKE_CUDA_COMPILER="/usr/local/cuda-10.2/bin/nvcc" \
-DCUDNN_VERSION="8.0" \
-DPROTOBUF_VERSION="3.8.0" \
-DTRT_OUT_DIR="`pwd`/out" \
-DCUDNN_LIB="/pdk_files/cudnn/usr/lib/aarch64-linux-gnu/"
Then make:
CC=aarch64-linux-gnu-gcc make -j8 |& tee test.log
Resulted with:
-- Up-to-date: /workspace/TensorRT/SHAKED/TensorRT/build/third_party.protobuf_aarch64/include/google/protobuf/compiler/plugin.proto
-- Up-to-date: /workspace/TensorRT/SHAKED/TensorRT/build/third_party.protobuf_aarch64/lib/cmake/protobuf/protobuf-targets.cmake
-- Installing: /workspace/TensorRT/SHAKED/TensorRT/build/third_party.protobuf_aarch64/lib/cmake/protobuf/protobuf-targets-release.cmake
-- Up-to-date: /workspace/TensorRT/SHAKED/TensorRT/build/third_party.protobuf_aarch64/lib/cmake/protobuf
-- Up-to-date: /workspace/TensorRT/SHAKED/TensorRT/build/third_party.protobuf_aarch64/lib/cmake/protobuf/protobuf-config.cmake
-- Up-to-date: /workspace/TensorRT/SHAKED/TensorRT/build/third_party.protobuf_aarch64/lib/cmake/protobuf/protobuf-config-version.cmake
-- Up-to-date: /workspace/TensorRT/SHAKED/TensorRT/build/third_party.protobuf_aarch64/lib/cmake/protobuf/protobuf-options.cmake
-- Up-to-date: /workspace/TensorRT/SHAKED/TensorRT/build/third_party.protobuf_aarch64/lib/cmake/protobuf/protobuf-module.cmake
[ 16%] Completed 'third_party.protobuf_aarch64'
[ 19%] Built target third_party.protobuf_aarch64
[ 25%] Built target onnx_proto
[ 25%] proto/trtcaffe.proto -> proto/trtcaffe.pb.cc proto/trtcaffe.pb.h
[ 26%] Linking CXX shared library ../../out/libnvonnxparser.so
[ 34%] Built target nvonnxparser_static
[ 35%] Linking CXX executable getSupportedAPITest
[ 37%] Linking CXX executable onnx2trt
[ 37%] Built target caffe_proto
Scanning dependencies of target nvcaffeparser_static
Scanning dependencies of target nvcaffeparser
[ 37%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/proto/trtcaffe.pb.cc.o
[ 38%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser.dir/caffeParser/opParsers/parseAbsVal.cpp.o
[ 40%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser.dir/proto/trtcaffe.pb.cc.o
[ 41%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/opParsers/parseAbsVal.cpp.o
[ 41%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser.dir/caffeParser/opParsers/parseBatchNorm.cpp.o
[ 47%] Built target nvonnxparser
[ 49%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser.dir/caffeParser/opParsers/parseBNLL.cpp.o
/usr/lib/gcc-cross/aarch64-linux-gnu/7/../../../../aarch64-linux-gnu/bin/ld: warning: libcudnn.so.8, needed by /pdk_files/tensorrt/lib/libnvinfer.so, not found (try using -rpath or -rpath-link)
/usr/lib/gcc-cross/aarch64-linux-gnu/7/../../../../aarch64-linux-gnu/bin/ld: warning: libcudart.so.10.2, needed by /pdk_files/tensorrt/lib/libnvinfer.so, not found (try using -rpath or -rpath-link)
/usr/lib/gcc-cross/aarch64-linux-gnu/7/../../../../aarch64-linux-gnu/bin/ld: warning: libnvrtc.so.10.2, needed by /pdk_files/tensorrt/lib/libnvinfer.so, not found (try using -rpath or -rpath-link)
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudaMemcpyAsync'
/pdk_files/tensorrt/lib/libnvinfer_plugin.so: undefined reference to `cudnnSetTensor4dDescriptor'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnPoolingForward@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnActivationForward@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnCreate@libcudnn.so.8'
/pdk_files/tensorrt/lib/libmyelin.so: undefined reference to `cudnnGetConvolutionBackwardFilterAlgorithmMaxCount@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudnnConvolutionBackwardData@libcudnn.so.8'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcGetProgramLogSize@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `nvrtcGetPTXSize@libnvrtc.so.10.2'
/pdk_files/tensorrt/lib/libnvinfer.so: undefined reference to `cudaStreamWaitEvent@libcudart.so.10.2'
...
...
...
[ 85%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/opParsers/parsePReLU.cpp.o
[ 85%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/opParsers/parseReduction.cpp.o
[ 86%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/opParsers/parseReLU.cpp.o
[ 86%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/opParsers/parseReshape.cpp.o
[ 88%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/opParsers/parseScale.cpp.o
[ 89%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/opParsers/parseSigmoid.cpp.o
[ 89%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/opParsers/parseSoftMax.cpp.o
[ 91%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/opParsers/parseTanH.cpp.o
[ 92%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeWeightFactory/caffeWeightFactory.cpp.o
[ 92%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/caffeParser/caffeParser.cpp.o
[ 94%] Building CXX object parsers/caffe/CMakeFiles/nvcaffeparser_static.dir/NvCaffeParser.cpp.o
[ 95%] Linking CXX shared library ../../out/libnvcaffeparser.so
[ 95%] Built target nvcaffeparser
[ 97%] Linking CXX static library ../../out/libnvcaffeparser_static.a
[ 98%] Built target nvcaffeparser_static
Makefile:151: recipe for target 'all' failed
make: *** [all] Error 2
full log (a.log):
@Shaked,
Your original problem was due to missing the -DCMAKE_TOOLCHAIN_FILE
argument. That is why you were running into incompatible ELF files. From looking at the CMake files I see you might need to set CUDNN_ROOT_DIR
to where cuDNN is located for AArch64. It should be the directory that contains lib(64)
and include
. Can you try building with make ... VERBOSE=1
which will give you more information including the -I
and -L
arguments that get generated. Then it would be good to compare those paths to the content of the directories using tree
or a similar command. Please attach the logs to the your comment if you have any output to share rather than paste the raw text into the comment because it's REALLY hard to follow the conversation now due to the large spans of text between comments.
I will close this issue since we already provide the solution, @Shaked please reopen if you still hit issues, thanks!
Reading through this github post helped me solve installing on a Jetson TX1 on JetPack 4.4.2 using 7.1.3 branch.
Commands Used:
> git clone -b 7.1.3 https://www.github.com/NVIDIA/TensorRT
> cd TensorRT
>git submodule update --init --recursive
>mkdir && cd build
> cmake .. -DTRT_LIB_DIB=/usr/lib/aarch64-linux-gnu -DTRT_OUT_DIR=/media/user/SD/TensorRT/build/out -DCMAKE_TOOLCHAIN=/media/user/SD/TensorRT/cmake/toolchains/cmake_aarch64.toolchain -DCUDA_VERSION=10.2
>$
>cc=aarch64-linux-gnu-gcc make -j8 |& tee test.log
Useful info: CMAKE_TOOLCHAIN may not be needed using these commands
-- Generating done
CMake Warning:
Manually-specified variables were not used by the project:
CMAKE_TOOLCHAIN
P.S. Toolchain may or may not be needed in cmake of TensorRT with these commands.
Many libraries needed are not found in the /targets directory of cuda. Libraries such as NVInfer....so are /usr/lib/aarch64-linux-gnu.
Hopefully this helps someone!
@Shaked and @ttyio
Have you solved the problem:
usr/lib/gcc-cross/aarch64-linux-gnu/7/../../../../aarch64-linux-gnu/bin/ld: warning: libcudnn.so.8, needed by /pdk_files/tensorrt/lib/libnvinfer.so, not found (try using -rpath or -rpath-link)
I used following cmake command (after fixing the dockerfile to build the dockercontainer and adding more log to cmakslist):
cmake .. -DTRT_LIB_DIR=$TRT_RELEASE/lib -DTRT_OUT_DIR=`pwd`/out -DBUILD_PLUGINS=OFF -DCMAKE_TOOLCHAIN_FILE=$TRT_SOURCE/cmake/toolchains/cmake_aarch64.toolchain -DCMAKE_CUDA_SEPARABLE_COMPILATION=OFF -DCUDA_VERSION=10.2 -DCUBLASLT_LIB="/usr/lib/aarch64-linux-gnu/libcublasLt.so" -DCUBLAS_LIB="/usr/lib/aarch64-linux-gnu/libcublas.so" -DCUDNN_LIB="/pdk_files/cudnn/lib/libcudnn.so.8"
root@0c622d0388ff:/workspace/TensorRT/build# cmake .. -DTRT_LIB_DIR=$TRT_RELEASE/lib -DTRT_OUT_DIR=`pwd`/out -DBUILD_PLUGINS=OFF -DCMAKE_TOOLCHAIN_FILE=$TRT_SOURCE/cmake/toolchains/cmake_aarch64.toolchain -DCMAKE_CUDA_SEPARABLE_COMPILATION=OFF -DCUDA_VERSION=10.2 -DCUBLASLT_LIB="/usr/lib/aarch64-linux-gnu/libcublasLt.so" -DCUBLAS_LIB="/usr/lib/aarch64-linux-gnu/libcublas.so" -DCUDNN_LIB="/pdk_files/cudnn/lib/libcudnn.so.8"
Building for TensorRT version: 7.2.1, library version: 7
-- The CXX compiler identification is GNU 7.5.0
-- The CUDA compiler identification is NVIDIA 10.2.89
-- Targeting TRT Platform: aarch64
-- CUDA version set to 10.2
-- cuDNN version set to 8.0
-- Protobuf version set to 3.0.0
-- Looking for C++ include pthread.h
-- Looking for C++ include pthread.h - found
-- Looking for pthread_create
-- Looking for pthread_create - not found
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE
-- Found PkgConfig: /usr/bin/pkg-config (found version "0.29.1")
-- Checking for one of the modules 'zlib'
-- Setting up another Protobuf build for cross compilation targeting aarch64-Linux
-- Using libprotobuf /workspace/TensorRT/build/third_party.protobuf_aarch64/lib/libprotobuf.a
-- found:/pdk_files/cudnn/lib/libcudnn.so.8
-- found:/usr/lib/aarch64-linux-gnu/libcublas.so
-- found:/usr/lib/aarch64-linux-gnu/libcublasLt.so
-- ========================= Importing and creating target nvinfer ==========================
-- Looking for library nvinfer
-- Library that was found /pdk_files/tensorrt/lib/libnvinfer.so
-- ==========================================================================================
-- ========================= Importing and creating target nvuffparser ==========================
-- Looking for library nvparsers
-- Library that was found /pdk_files/tensorrt/lib/libnvparsers.so
-- ==========================================================================================
-- found:/usr/local/cuda-10.2/targets/aarch64-linux/lib/libcudart.so
CMake Warning at CMakeLists.txt:181 (message):
Detected CUDA version is < 11.0. SM80 not supported.
-- GPU_ARCHS is not defined. Generating CUDA code for default SMs: 35;53;61;70;75;72
-- ========================= Importing and creating target nvinfer_plugin ==========================
-- Looking for library nvinfer_plugin
-- Library that was found /pdk_files/tensorrt/lib/libnvinfer_plugin.so
-- ==========================================================================================
-- Protobuf proto/trtcaffe.proto -> proto/trtcaffe.pb.cc proto/trtcaffe.pb.h
-- /workspace/TensorRT/build/parsers/caffe
-- The C compiler identification is GNU 7.5.0
-- Build type not set - defaulting to Release
Generated: /workspace/TensorRT/build/parsers/onnx/third_party/onnx/onnx/onnx_onnx2trt_onnx-ml.proto
Generated: /workspace/TensorRT/build/parsers/onnx/third_party/onnx/onnx/onnx-operators_onnx2trt_onnx-ml.proto
--
-- ******** Summary ********
-- CMake version : 3.14.4
-- CMake command : /usr/local/bin/cmake
-- System : Linux
-- C++ compiler : /usr/bin/aarch64-linux-gnu-g++
-- C++ compiler version : 7.5.0
-- CXX flags : -Wno-deprecated-declarations -DBUILD_SYSTEM=cmake_oss -Wall -Wno-deprecated-declarations -Wno-unused-function -Wnon-virtual-dtor
-- Build type : Release
-- Compile definitions : _PROTOBUF_INSTALL_DIR=/workspace/TensorRT/build/third_party.protobuf;ONNX_NAMESPACE=onnx2trt_onnx
-- CMAKE_PREFIX_PATH :
-- CMAKE_INSTALL_PREFIX : /pdk_files/tensorrt/lib/..
-- CMAKE_MODULE_PATH :
--
-- ONNX version : 1.6.0
-- ONNX NAMESPACE : onnx2trt_onnx
-- ONNX_BUILD_TESTS : OFF
-- ONNX_BUILD_BENCHMARKS : OFF
-- ONNX_USE_LITE_PROTO : OFF
-- ONNXIFI_DUMMY_BACKEND : OFF
-- ONNXIFI_ENABLE_EXT : OFF
--
-- Protobuf compiler :
-- Protobuf includes :
-- Protobuf libraries :
-- BUILD_ONNX_PYTHON : OFF
-- Found TensorRT headers at /workspace/TensorRT/include
-- Find TensorRT libs at /pdk_files/tensorrt/lib/libnvinfer.so;/pdk_files/tensorrt/lib/libnvinfer_plugin.so;/pdk_files/tensorrt/lib/libmyelin.so
-- Found TENSORRT: /workspace/TensorRT/include
-- Adding new sample: sample_algorithm_selector
-- - Parsers Used: caffe
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_char_rnn
-- - Parsers Used: uff;caffe;onnx
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_dynamic_reshape
-- - Parsers Used: onnx
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_fasterRCNN
-- - Parsers Used: caffe
-- - InferPlugin Used: ON
-- - Licensing: opensource
-- Adding new sample: sample_googlenet
-- - Parsers Used: caffe
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_int8
-- - Parsers Used: caffe
-- - InferPlugin Used: ON
-- - Licensing: opensource
-- Adding new sample: sample_int8_api
-- - Parsers Used: onnx
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_mlp
-- - Parsers Used: caffe
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_mnist
-- - Parsers Used: caffe
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_mnist_api
-- - Parsers Used: caffe
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_movielens
-- - Parsers Used: uff
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_movielens_mps
-- - Parsers Used: uff
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_nmt
-- - Parsers Used: none
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_onnx_mnist
-- - Parsers Used: onnx
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_plugin
-- - Parsers Used: caffe
-- - InferPlugin Used: ON
-- - Licensing: opensource
-- Adding new sample: sample_reformat_free_io
-- - Parsers Used: caffe
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_ssd
-- - Parsers Used: caffe
-- - InferPlugin Used: ON
-- - Licensing: opensource
-- Adding new sample: sample_uff_fasterRCNN
-- - Parsers Used: uff
-- - InferPlugin Used: ON
-- - Licensing: opensource
-- Adding new sample: sample_uff_maskRCNN
-- - Parsers Used: uff
-- - InferPlugin Used: ON
-- - Licensing: opensource
-- Adding new sample: sample_uff_mnist
-- - Parsers Used: uff
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_uff_plugin_v2_ext
-- - Parsers Used: uff
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Adding new sample: sample_uff_ssd
-- - Parsers Used: uff
-- - InferPlugin Used: ON
-- - Licensing: opensource
-- Adding new sample: sample_onnx_mnist_coord_conv_ac
-- - Parsers Used: onnx
-- - InferPlugin Used: ON
-- - Licensing: opensource
-- Adding new sample: trtexec
-- - Parsers Used: caffe;uff;onnx
-- - InferPlugin Used: OFF
-- - Licensing: opensource
-- Configuring done
-- Generating done
-- Build files have been written to: /workspace/TensorRT/build
root@0c622d0388ff:/workspace/TensorRT/build#
I can add more information if required. I think there is just some linker information missing, but i am not sure how to correct adapt the cmake files...
Hello,
I'm trying to compile this repository for ARM. This is the steps I have taken so far:
sdkmanager
:ubuntu-cross-aarch64.Dockerfile
image. Currently without GPUs as I'm saving costs.sudo cmake .. -DDTRT_LIB_DIR=/pdk_files/tensorrt/usr/lib/aarch64-linux-gnu/ -DTRT_OUT_DIR=
pwd/out -DCUDNN_LIB=/usr/local/cuda/lib64 -DTENSORRT_LIBRARY_INFER=/pdk_files/tensorrt/lib/ -DTENSORRT_LIBRARY_INFER_PLUGIN=/pdk_files/tensorrt/lib -DTENSORRT_LIBRARY_MYELIN=/pdk_files/tensorrt/lib
sudo make -j$(nproc)
Also fails as it wants the following to in
/usr/include
, which I symlinked:Fix cmake wrong param and rerun:
trtuser@986425528e26:/workspace/TensorRT/build$ sudo cmake .. -DTRT_LIB_DIR=/pdk_files/tensorrt/lib/stubs/ -DTRT_OUT_DIR=
pwd/out -DCUDNN_LIB=/usr/local/cuda/lib64 -DTENSORRT_LIBRARY_INFER=/pdk_files/tensorrt/lib/stubs/ -DTENSORRT_LIBRARY_INFER_PLUGIN=/pdk_files/tensorrt/lib/stubs/ -DTENSORRT_LIBRARY_MYELIN=/pdk_files/tensorrt/lib/
Try
make -j$(nproc)
:Not sure what to do from here. I was thinking maybe cmake/make needs a flag to use the right g++/gcc same as with the stub, so I tried also adding the following code taken from this post:
But this complains on a different issue:
Even though these variables are defined within the cmake execution.
My main goal is to build TensorRT components on a x86 VM for ARM and then use it on:
Did I understand this completely wrong? Is this even possible?
Thank you! Shaked
UPDATE 1:
Following https://github.com/NVIDIA/TensorRT/issues/643 and https://github.com/NVIDIA/TensorRT/issues/9#issuecomment-575106079, I have ran:
Then
UPDATE 2
I figured that I had placed a non existing path for
libcublas.so
. I also tried to use the/pdk_files/tensorrt/lib/
path instead of/pdk_files/tensorrt/lib/stubs
, this is what I ended up with:Returns:
Then
Returns