Tensorrt: Build tensorRT OSS successful but do not have any symbol plugin in library file.

Created on 11 Nov 2020 · 6Comments · Source: NVIDIA/TensorRT

Description

Hello I follow build TensorRT from source, but maybe have something miss take need help:

Check symbol plugin exist with command

nm -gDC file/libnvinfer_plugin.so

Bellow the step I build source.

add some PATH default:

export TRT_SOURCE=`pwd`
export TRT_RELEASE=`pwd`/release

make build folder and do cmake file

mkdir build && cd build

cmake .. -DTRT_LIB_DIR=$TRT_RELEASE/lib -DTRT_OUT_DIR=`pwd`/out -DBUILD_PLUGINS=ON -DBUILD_SAMPLES=OFF -DBUILD_PARSERS=OFF -DCUDA_VERSION=10.2

make -j12

output here:

```{r, attr.output='style="max-height: 100px;"'}
Building for TensorRT version: 7.1.3, library version: 7
-- Targeting TRT Platform: x86_64
-- CUDA version set to 10.2
-- cuDNN version set to 8.0
-- Protobuf version set to 3.0.0
-- ========================= Importing and creating target nvinfer ==========================
-- Looking for library nvinfer
-- Library that was found /usr/lib/x86_64-linux-gnu/libnvinfer.so
-- ==========================================================================================
-- ========================= Importing and creating target nvuffparser ==========================
-- Looking for library nvparsers
-- Library that was found /usr/lib/x86_64-linux-gnu/libnvparsers.so
-- ==========================================================================================
CMake Warning at CMakeLists.txt:157 (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 nvcaffeparser ==========================
-- Looking for library nvparsers
-- Library that was found /usr/lib/x86_64-linux-gnu/libnvparsers.so
-- ==========================================================================================
-- ========================= Importing and creating target nvonnxparser ==========================
-- Looking for library nvonnxparser
-- Library that was found /usr/lib/x86_64-linux-gnu/libnvonnxparser.so
-- ==========================================================================================
-- Configuring done
-- Generating done
-- Build files have been written to: /workspace/tensorRT/version-7.1/TensorRT/build

```{r, attr.output='style="max-height: 100px;"'}
[  0%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/nmsPlugin/nmsPlugin.cpp.o
[  0%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/reorgPlugin/reorgPlugin.cpp.o
[  1%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/batchedNMSPlugin/batchedNMSPlugin.cpp.o
[  2%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/gridAnchorPlugin/gridAnchorPlugin.cpp.o
[  2%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/nvFasterRCNN/nvFasterRCNNPlugin.cpp.o
[  4%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/normalizePlugin/normalizePlugin.cpp.o
[  4%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/regionPlugin/regionPlugin.cpp.o
[  5%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/priorBoxPlugin/priorBoxPlugin.cpp.o
[  5%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/nmsPlugin/nmsPlugin.cpp.o
[  6%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/priorBoxPlugin/priorBoxPlugin.cpp.o
[  7%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/flattenConcat/flattenConcat.cpp.o
[  8%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/normalizePlugin/normalizePlugin.cpp.o
[  8%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/cropAndResizePlugin/cropAndResizePlugin.cpp.o
[  9%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/proposalPlugin/proposalPlugin.cpp.o
[  9%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/reorgPlugin/reorgPlugin.cpp.o
[ 10%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/batchTilePlugin/batchTilePlugin.cpp.o
[ 10%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/detectionLayerPlugin/detectionLayerPlugin.cpp.o
[ 11%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/gridAnchorPlugin/gridAnchorPlugin.cpp.o
[ 12%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/proposalLayerPlugin/proposalLayerPlugin.cpp.o
[ 13%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/pyramidROIAlignPlugin/pyramidROIAlignPlugin.cpp.o
[ 14%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/regionPlugin/regionPlugin.cpp.o
[ 14%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/generateDetectionPlugin/generateDetectionPlugin.cpp.o
[ 15%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/multilevelProposeROI/multilevelProposeROIPlugin.cpp.o
[ 16%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/multilevelCropAndResizePlugin/multilevelCropAndResizePlugin.cpp.o
[ 16%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/resizeNearestPlugin/resizeNearestPlugin.cpp.o
[ 17%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/specialSlicePlugin/specialSlicePlugin.cpp.o
[ 17%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/nvFasterRCNN/nvFasterRCNNPlugin.cpp.o
[ 18%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/batchedNMSPlugin/batchedNMSPlugin.cpp.o
[ 19%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/instanceNormalizationPlugin/instanceNormalizationPlugin.cpp.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/multilevelProposeROI/multilevelProposeROIPlugin.cpp: In member function 'virtual int nvinfer1::plugin::MultilevelProposeROI::enqueue(int, const void* const*, void**, void*, cudaStream_t)':
/workspace/tensorRT/version-7.1/TensorRT/plugin/multilevelProposeROI/multilevelProposeROIPlugin.cpp:395:44: warning: pointer of type 'void *' used in arithmetic [-Wpointer-arith]
             mParam, proposal_ws, workspace + kernel_workspace_offset,
                                  ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~
/workspace/tensorRT/version-7.1/TensorRT/plugin/multilevelProposeROI/multilevelProposeROIPlugin.cpp:408:19: warning: pointer of type 'void *' used in arithmetic [-Wpointer-arith]
         workspace + kernel_workspace_offset, ctopk_ws, reinterpret_cast<void**>(mDeviceScores),
         ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~
[ 19%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/groupNormalizationPlugin/groupNormalizationKernel.cu.o
[ 20%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/flattenConcat/flattenConcat.cpp.o
[ 21%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/groupNormalizationPlugin/groupNormalizationPlugin.cpp.o
[ 22%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/coordConvACPlugin/coordConvACPlugin.cpp.o
[ 23%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/fcPlugin/fcPlugin.cpp.o
[ 23%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/embLayerNormPlugin/embLayerNormPlugin.cpp.o
[ 23%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/cropAndResizePlugin/cropAndResizePlugin.cpp.o
[ 24%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/geluPlugin/geluPlugin.cpp.o
[ 25%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/proposalPlugin/proposalPlugin.cpp.o
[ 25%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_128_64_kernel.sm75.cpp.o
[ 26%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/batchTilePlugin/batchTilePlugin.cpp.o
[ 26%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/detectionLayerPlugin/detectionLayerPlugin.cpp.o
[ 27%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_128_64_kernel.sm80.cpp.o
[ 28%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/proposalLayerPlugin/proposalLayerPlugin.cpp.o
[ 29%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_384_64_kernel.sm75.cpp.o
[ 29%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_384_64_kernel.sm80.cpp.o
[ 30%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_128_64_kernel.sm75.cpp.o
[ 31%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_128_64_kernel.sm80.cpp.o
[ 32%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/pyramidROIAlignPlugin/pyramidROIAlignPlugin.cpp.o
[ 32%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_384_64_kernel.sm75.cpp.o
[ 32%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/generateDetectionPlugin/generateDetectionPlugin.cpp.o
[ 33%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_384_64_kernel.sm80.cpp.o
[ 34%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/multilevelProposeROI/multilevelProposeROIPlugin.cpp.o
[ 35%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/qkvToContextPlugin.cpp.o
[ 35%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/skipLayerNormPlugin/skipLayerNormPlugin.cpp.o
[ 36%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/kernel.cpp.o
[ 37%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/multilevelCropAndResizePlugin/multilevelCropAndResizePlugin.cpp.o
[ 38%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/common/checkMacrosPlugin.cpp.o
[ 38%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/resizeNearestPlugin/resizeNearestPlugin.cpp.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/multilevelProposeROI/multilevelProposeROIPlugin.cpp: In member function 'virtual int nvinfer1::plugin::MultilevelProposeROI::enqueue(int, const void* const*, void**, void*, cudaStream_t)':
/workspace/tensorRT/version-7.1/TensorRT/plugin/multilevelProposeROI/multilevelProposeROIPlugin.cpp:395:44: warning: pointer of type 'void *' used in arithmetic [-Wpointer-arith]
             mParam, proposal_ws, workspace + kernel_workspace_offset,
                                  ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~
/workspace/tensorRT/version-7.1/TensorRT/plugin/multilevelProposeROI/multilevelProposeROIPlugin.cpp:408:19: warning: pointer of type 'void *' used in arithmetic [-Wpointer-arith]
         workspace + kernel_workspace_offset, ctopk_ws, reinterpret_cast<void**>(mDeviceScores),
         ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~
[ 39%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/specialSlicePlugin/specialSlicePlugin.cpp.o
[ 39%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/common/cudaDriverWrapper.cpp.o
[ 40%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/common/nmsHelper.cpp.o
[ 41%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/common/reducedMathPlugin.cpp.o
[ 42%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/instanceNormalizationPlugin/instanceNormalizationPlugin.cpp.o
[ 42%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/batchedNMSPlugin/batchedNMSInference.cu.o
[ 43%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/batchedNMSPlugin/gatherNMSOutputs.cu.o
[ 43%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/groupNormalizationPlugin/groupNormalizationKernel.cu.o
[ 44%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/coordConvACPlugin/coordConvACPlugin.cu.o
[ 44%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/allClassNMS.cu.o
[ 45%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/groupNormalizationPlugin/groupNormalizationPlugin.cpp.o
[ 46%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/bboxDeltas2Proposals.cu.o
[ 47%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/coordConvACPlugin/coordConvACPlugin.cpp.o
[ 48%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/common.cu.o
[ 48%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/cropAndResizeKernel.cu.o
[ 48%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/embLayerNormPlugin/embLayerNormPlugin.cpp.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/batchedNMSPlugin/gatherNMSOutputs.cu(82): warning: argument is incompatible with corresponding format string conversion
          detected during:
            instantiation of "void gatherNMSOutputs_kernel<T_BBOX,T_SCORE,nthds_per_cta>(__nv_bool, int, int, int, int, int, const int *, const T_SCORE *, const T_BBOX *, int *, T_BBOX *, T_BBOX *, T_BBOX *, __nv_bool) [with T_BBOX=float, T_SCORE=float, nthds_per_cta=32U]" 
(117): here
            instantiation of "pluginStatus_t gatherNMSOutputs_gpu<T_BBOX,T_SCORE>(cudaStream_t, __nv_bool, int, int, int, int, int, const void *, const void *, const void *, void *, void *, void *, void *, __nv_bool) [with T_BBOX=float, T_SCORE=float]" 
(165): here

[ 49%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/fcPlugin/fcPlugin.cpp.o
[ 50%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/decodeBBoxes.cu.o
[ 51%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/geluPlugin/geluPlugin.cpp.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/batchedNMSPlugin/gatherNMSOutputs.cu(82): warning: argument is incompatible with corresponding format string conversion
          detected during:
            instantiation of "void gatherNMSOutputs_kernel<T_BBOX,T_SCORE,nthds_per_cta>(__nv_bool, int, int, int, int, int, const int *, const T_SCORE *, const T_BBOX *, int *, T_BBOX *, T_BBOX *, T_BBOX *, __nv_bool) [with T_BBOX=float, T_SCORE=float, nthds_per_cta=32U]" 
(117): here
            instantiation of "pluginStatus_t gatherNMSOutputs_gpu<T_BBOX,T_SCORE>(cudaStream_t, __nv_bool, int, int, int, int, int, const void *, const void *, const void *, void *, void *, void *, void *, __nv_bool) [with T_BBOX=float, T_SCORE=float]" 
(165): here

/workspace/tensorRT/version-7.1/TensorRT/plugin/batchedNMSPlugin/gatherNMSOutputs.cu(82): warning: argument is incompatible with corresponding format string conversion
          detected during:
            instantiation of "void gatherNMSOutputs_kernel<T_BBOX,T_SCORE,nthds_per_cta>(__nv_bool, int, int, int, int, int, const int *, const T_SCORE *, const T_BBOX *, int *, T_BBOX *, T_BBOX *, T_BBOX *, __nv_bool) [with T_BBOX=float, T_SCORE=float, nthds_per_cta=32U]" 
(117): here
            instantiation of "pluginStatus_t gatherNMSOutputs_gpu<T_BBOX,T_SCORE>(cudaStream_t, __nv_bool, int, int, int, int, int, const void *, const void *, const void *, void *, void *, void *, void *, __nv_bool) [with T_BBOX=float, T_SCORE=float]" 
(165): here

[ 51%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_128_64_kernel.sm75.cpp.o
[ 52%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_128_64_kernel.sm80.cpp.o
[ 53%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_384_64_kernel.sm75.cpp.o
[ 53%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_384_64_kernel.sm80.cpp.o
[ 54%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_128_64_kernel.sm75.cpp.o
[ 55%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_128_64_kernel.sm80.cpp.o
[ 55%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_384_64_kernel.sm75.cpp.o
[ 56%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_384_64_kernel.sm80.cpp.o
[ 57%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/bertQKVToContextPlugin/qkvToContextPlugin.cpp.o
[ 57%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/skipLayerNormPlugin/skipLayerNormPlugin.cpp.o
[ 58%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/kernel.cpp.o
[ 59%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/checkMacrosPlugin.cpp.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/batchedNMSPlugin/gatherNMSOutputs.cu(82): warning: argument is incompatible with corresponding format string conversion
          detected during:
            instantiation of "void gatherNMSOutputs_kernel<T_BBOX,T_SCORE,nthds_per_cta>(__nv_bool, int, int, int, int, int, const int *, const T_SCORE *, const T_BBOX *, int *, T_BBOX *, T_BBOX *, T_BBOX *, __nv_bool) [with T_BBOX=float, T_SCORE=float, nthds_per_cta=32U]" 
(117): here
            instantiation of "pluginStatus_t gatherNMSOutputs_gpu<T_BBOX,T_SCORE>(cudaStream_t, __nv_bool, int, int, int, int, int, const void *, const void *, const void *, void *, void *, void *, void *, __nv_bool) [with T_BBOX=float, T_SCORE=float]" 
(165): here

[ 59%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/cudaDriverWrapper.cpp.o
[ 60%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/detectionForward.cu.o
[ 61%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/nmsHelper.cpp.o
[ 62%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/reducedMathPlugin.cpp.o
[ 62%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/batchedNMSPlugin/batchedNMSInference.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/batchedNMSPlugin/gatherNMSOutputs.cu(82): warning: argument is incompatible with corresponding format string conversion
          detected during:
            instantiation of "void gatherNMSOutputs_kernel<T_BBOX,T_SCORE,nthds_per_cta>(__nv_bool, int, int, int, int, int, const int *, const T_SCORE *, const T_BBOX *, int *, T_BBOX *, T_BBOX *, T_BBOX *, __nv_bool) [with T_BBOX=float, T_SCORE=float, nthds_per_cta=32U]" 
(117): here
            instantiation of "pluginStatus_t gatherNMSOutputs_gpu<T_BBOX,T_SCORE>(cudaStream_t, __nv_bool, int, int, int, int, int, const void *, const void *, const void *, void *, void *, void *, void *, __nv_bool) [with T_BBOX=float, T_SCORE=float]" 
(165): here

[ 63%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/batchedNMSPlugin/gatherNMSOutputs.cu.o
[ 64%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/coordConvACPlugin/coordConvACPlugin.cu.o
[ 64%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/allClassNMS.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/batchedNMSPlugin/gatherNMSOutputs.cu(82): warning: argument is incompatible with corresponding format string conversion
          detected during:
            instantiation of "void gatherNMSOutputs_kernel<T_BBOX,T_SCORE,nthds_per_cta>(__nv_bool, int, int, int, int, int, const int *, const T_SCORE *, const T_BBOX *, int *, T_BBOX *, T_BBOX *, T_BBOX *, __nv_bool) [with T_BBOX=float, T_SCORE=float, nthds_per_cta=32U]" 
(117): here
            instantiation of "pluginStatus_t gatherNMSOutputs_gpu<T_BBOX,T_SCORE>(cudaStream_t, __nv_bool, int, int, int, int, int, const void *, const void *, const void *, void *, void *, void *, void *, __nv_bool) [with T_BBOX=float, T_SCORE=float]" 
(165): here

[ 64%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/extractFgScores.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/batchedNMSPlugin/gatherNMSOutputs.cu(82): warning: argument is incompatible with corresponding format string conversion
          detected during:
            instantiation of "void gatherNMSOutputs_kernel<T_BBOX,T_SCORE,nthds_per_cta>(__nv_bool, int, int, int, int, int, const int *, const T_SCORE *, const T_BBOX *, int *, T_BBOX *, T_BBOX *, T_BBOX *, __nv_bool) [with T_BBOX=float, T_SCORE=float, nthds_per_cta=32U]" 
(117): here
            instantiation of "pluginStatus_t gatherNMSOutputs_gpu<T_BBOX,T_SCORE>(cudaStream_t, __nv_bool, int, int, int, int, int, const void *, const void *, const void *, void *, void *, void *, void *, __nv_bool) [with T_BBOX=float, T_SCORE=float]" 
(165): here

[ 65%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/gatherTopDetections.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/batchedNMSPlugin/gatherNMSOutputs.cu(82): warning: argument is incompatible with corresponding format string conversion
          detected during:
            instantiation of "void gatherNMSOutputs_kernel<T_BBOX,T_SCORE,nthds_per_cta>(__nv_bool, int, int, int, int, int, const int *, const T_SCORE *, const T_BBOX *, int *, T_BBOX *, T_BBOX *, T_BBOX *, __nv_bool) [with T_BBOX=float, T_SCORE=float, nthds_per_cta=32U]" 
(117): here
            instantiation of "pluginStatus_t gatherNMSOutputs_gpu<T_BBOX,T_SCORE>(cudaStream_t, __nv_bool, int, int, int, int, int, const void *, const void *, const void *, void *, void *, void *, void *, __nv_bool) [with T_BBOX=float, T_SCORE=float]" 
(165): here

[ 66%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/generateAnchors.cu.o
[ 66%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/gridAnchorLayer.cu.o
[ 67%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/maskRCNNKernels.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/batchedNMSPlugin/gatherNMSOutputs.cu(82): warning: argument is incompatible with corresponding format string conversion
          detected during:
            instantiation of "void gatherNMSOutputs_kernel<T_BBOX,T_SCORE,nthds_per_cta>(__nv_bool, int, int, int, int, int, const int *, const T_SCORE *, const T_BBOX *, int *, T_BBOX *, T_BBOX *, T_BBOX *, __nv_bool) [with T_BBOX=float, T_SCORE=float, nthds_per_cta=32U]" 
(117): here
            instantiation of "pluginStatus_t gatherNMSOutputs_gpu<T_BBOX,T_SCORE>(cudaStream_t, __nv_bool, int, int, int, int, int, const void *, const void *, const void *, void *, void *, void *, void *, __nv_bool) [with T_BBOX=float, T_SCORE=float]" 
(165): here

/workspace/tensorRT/version-7.1/TensorRT/plugin/batchedNMSPlugin/gatherNMSOutputs.cu(82): warning: argument is incompatible with corresponding format string conversion
          detected during:
            instantiation of "void gatherNMSOutputs_kernel<T_BBOX,T_SCORE,nthds_per_cta>(__nv_bool, int, int, int, int, int, const int *, const T_SCORE *, const T_BBOX *, int *, T_BBOX *, T_BBOX *, T_BBOX *, __nv_bool) [with T_BBOX=float, T_SCORE=float, nthds_per_cta=32U]" 
(117): here
            instantiation of "pluginStatus_t gatherNMSOutputs_gpu<T_BBOX,T_SCORE>(cudaStream_t, __nv_bool, int, int, int, int, int, const void *, const void *, const void *, void *, void *, void *, void *, __nv_bool) [with T_BBOX=float, T_SCORE=float]" 
(165): here

[ 68%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/nmsLayer.cu.o
[ 69%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/bboxDeltas2Proposals.cu.o
[ 70%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/common.cu.o
[ 70%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/normalizeLayer.cu.o
[ 71%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/permuteData.cu.o
[ 72%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/priorBoxLayer.cu.o
[ 72%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/proposalKernel.cu.o
[ 73%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/proposalsForward.cu.o
[ 74%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/regionForward.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/common/kernels/proposalKernel.cu(34): warning: variable "ALIGNMENT" was declared but never referenced

[ 74%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/reorgForward.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/common/kernels/proposalKernel.cu(34): warning: variable "ALIGNMENT" was declared but never referenced

/workspace/tensorRT/version-7.1/TensorRT/plugin/common/kernels/proposalKernel.cu(34): warning: variable "ALIGNMENT" was declared but never referenced

[ 74%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/cropAndResizeKernel.cu.o
[ 75%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/decodeBBoxes.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/common/kernels/proposalKernel.cu(34): warning: variable "ALIGNMENT" was declared but never referenced

[ 76%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/roiPooling.cu.o
[ 77%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/rproiInferenceFused.cu.o
[ 78%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/detectionForward.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/common/kernels/proposalKernel.cu(34): warning: variable "ALIGNMENT" was declared but never referenced

[ 78%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/extractFgScores.cu.o
[ 79%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/gatherTopDetections.cu.o
[ 79%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/sortScoresPerClass.cu.o
[ 80%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/common/kernels/sortScoresPerImage.cu.o
[ 81%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/embLayerNormPlugin/embLayerNormKernel.cu.o
[ 81%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/geluPlugin/geluKernel.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/embLayerNormPlugin/embLayerNormKernel.cu(51): warning: variable "warp_m" was declared but never referenced

/workspace/tensorRT/version-7.1/TensorRT/plugin/embLayerNormPlugin/embLayerNormKernel.cu(51): warning: variable "warp_m" was declared but never referenced

[ 82%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/generateAnchors.cu.o
[ 83%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/qkvToContext.cu.o
[ 84%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin.dir/skipLayerNormPlugin/skipLayerNormKernel.cu.o
[ 84%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/gridAnchorLayer.cu.o
[ 85%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/maskRCNNKernels.cu.o
[ 86%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/nmsLayer.cu.o
[ 86%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/InferPlugin.cpp.o
[ 87%] Building CXX object plugin/CMakeFiles/nvinfer_plugin.dir/__/samples/common/logger.cpp.o
[ 87%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/normalizeLayer.cu.o
[ 88%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/permuteData.cu.o
[ 89%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/priorBoxLayer.cu.o
[ 89%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/proposalKernel.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/common/kernels/proposalKernel.cu(34): warning: variable "ALIGNMENT" was declared but never referenced

[ 90%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/proposalsForward.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/common/kernels/proposalKernel.cu(34): warning: variable "ALIGNMENT" was declared but never referenced

/workspace/tensorRT/version-7.1/TensorRT/plugin/common/kernels/proposalKernel.cu(34): warning: variable "ALIGNMENT" was declared but never referenced

[ 91%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/regionForward.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/common/kernels/proposalKernel.cu(34): warning: variable "ALIGNMENT" was declared but never referenced

[ 91%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/reorgForward.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/common/kernels/proposalKernel.cu(34): warning: variable "ALIGNMENT" was declared but never referenced

[ 92%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/roiPooling.cu.o
[ 93%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/rproiInferenceFused.cu.o
[ 93%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/sortScoresPerClass.cu.o
[ 94%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/common/kernels/sortScoresPerImage.cu.o
[ 95%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/embLayerNormPlugin/embLayerNormKernel.cu.o
[ 95%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/geluPlugin/geluKernel.cu.o
[ 96%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/bertQKVToContextPlugin/qkvToContext.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/embLayerNormPlugin/embLayerNormKernel.cu(51): warning: variable "warp_m" was declared but never referenced

[ 97%] Building CUDA object plugin/CMakeFiles/nvinfer_plugin_static.dir/skipLayerNormPlugin/skipLayerNormKernel.cu.o
/workspace/tensorRT/version-7.1/TensorRT/plugin/embLayerNormPlugin/embLayerNormKernel.cu(51): warning: variable "warp_m" was declared but never referenced

[ 97%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/InferPlugin.cpp.o
[ 98%] Building CXX object plugin/CMakeFiles/nvinfer_plugin_static.dir/__/samples/common/logger.cpp.o
[ 99%] Linking CXX shared library ../out/libnvinfer_plugin.so
[ 99%] Built target nvinfer_plugin
[100%] Linking CXX static library ../out/libnvinfer_plugin_static.a
[100%] Built target nvinfer_plugin_static

Environment

TensorRT Version: 7.1
GPU Type: 1080i
Nvidia Driver Version: 450.66
CUDA Version: 10.2
CUDNN Version: 8.0

Relevant Files

check out: my-libnvinfer_plugin.so and origin-libnvinfer_plugin.so
https://www.dropbox.com/sh/zpdvxjc673spqaj/AAAcbV-_F29SOdYnoZ9XqUHha?dl=0

OSS Build triaged

Source

ledinhtri97

All 6 comments

Nothing jumps out from the build log at least. Can you try building with VERBOSE=1 to make sure that the objects are all being linked into the library?

CC @rajeevsrao

pranavm-nvidia on 11 Nov 2020

here is the last log when add VERBOSE=1.

[100%] Linking CXX shared library ../out/libnvinfer_plugin.so
cd /workspace/tensorRT/version-7.1/TensorRT/build/plugin && /usr/local/bin/cmake -E cmake_link_script CMakeFiles/nvinfer_plugin.dir/link.txt --verbose=1
/usr/bin/g++ -fPIC -Wno-deprecated-declarations  -DBUILD_SYSTEM=cmake_oss -O3 -DNDEBUG -Wl,--exclude-libs,ALL -Wl,--version-script=/workspace/tensorRT/version-7.1/TensorRT/plugin/exports.map -Wl,--no-undefined -shared -Wl,-soname,libnvinfer_plugin.so.7 -o ../out/libnvinfer_plugin.so.7.1.3 CMakeFiles/nvinfer_plugin.dir/nmsPlugin/nmsPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/normalizePlugin/normalizePlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/priorBoxPlugin/priorBoxPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/reorgPlugin/reorgPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/gridAnchorPlugin/gridAnchorPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/regionPlugin/regionPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/nvFasterRCNN/nvFasterRCNNPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/batchedNMSPlugin/batchedNMSPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/flattenConcat/flattenConcat.cpp.o CMakeFiles/nvinfer_plugin.dir/cropAndResizePlugin/cropAndResizePlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/proposalPlugin/proposalPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/batchTilePlugin/batchTilePlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/detectionLayerPlugin/detectionLayerPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/proposalLayerPlugin/proposalLayerPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/pyramidROIAlignPlugin/pyramidROIAlignPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/generateDetectionPlugin/generateDetectionPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/multilevelProposeROI/multilevelProposeROIPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/multilevelCropAndResizePlugin/multilevelCropAndResizePlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/resizeNearestPlugin/resizeNearestPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/specialSlicePlugin/specialSlicePlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/instanceNormalizationPlugin/instanceNormalizationPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/groupNormalizationPlugin/groupNormalizationKernel.cu.o CMakeFiles/nvinfer_plugin.dir/groupNormalizationPlugin/groupNormalizationPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/coordConvACPlugin/coordConvACPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/embLayerNormPlugin/embLayerNormPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/fcPlugin/fcPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/geluPlugin/geluPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_128_64_kernel.sm75.cpp.o CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_128_64_kernel.sm80.cpp.o CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_384_64_kernel.sm75.cpp.o CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_fp16_384_64_kernel.sm80.cpp.o CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_128_64_kernel.sm75.cpp.o CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_128_64_kernel.sm80.cpp.o CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_384_64_kernel.sm75.cpp.o CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/fused_multihead_attention_int8_384_64_kernel.sm80.cpp.o CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/qkvToContextPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/skipLayerNormPlugin/skipLayerNormPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/common/kernels/kernel.cpp.o CMakeFiles/nvinfer_plugin.dir/common/checkMacrosPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/common/cudaDriverWrapper.cpp.o CMakeFiles/nvinfer_plugin.dir/common/nmsHelper.cpp.o CMakeFiles/nvinfer_plugin.dir/common/reducedMathPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/batchedNMSPlugin/batchedNMSInference.cu.o CMakeFiles/nvinfer_plugin.dir/batchedNMSPlugin/gatherNMSOutputs.cu.o CMakeFiles/nvinfer_plugin.dir/coordConvACPlugin/coordConvACPlugin.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/allClassNMS.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/bboxDeltas2Proposals.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/common.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/cropAndResizeKernel.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/decodeBBoxes.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/detectionForward.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/extractFgScores.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/gatherTopDetections.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/generateAnchors.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/gridAnchorLayer.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/maskRCNNKernels.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/nmsLayer.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/normalizeLayer.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/permuteData.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/priorBoxLayer.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/proposalKernel.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/proposalsForward.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/regionForward.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/reorgForward.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/roiPooling.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/rproiInferenceFused.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/sortScoresPerClass.cu.o CMakeFiles/nvinfer_plugin.dir/common/kernels/sortScoresPerImage.cu.o CMakeFiles/nvinfer_plugin.dir/embLayerNormPlugin/embLayerNormKernel.cu.o CMakeFiles/nvinfer_plugin.dir/geluPlugin/geluKernel.cu.o CMakeFiles/nvinfer_plugin.dir/bertQKVToContextPlugin/qkvToContext.cu.o CMakeFiles/nvinfer_plugin.dir/skipLayerNormPlugin/skipLayerNormKernel.cu.o CMakeFiles/nvinfer_plugin.dir/InferPlugin.cpp.o CMakeFiles/nvinfer_plugin.dir/__/samples/common/logger.cpp.o   -L/usr/local/cuda/targets/x86_64-linux/lib  /usr/lib/x86_64-linux-gnu/libcublas.so /usr/lib/x86_64-linux-gnu/libcublasLt.so /usr/local/cuda/lib64/libcudart.so /usr/lib/x86_64-linux-gnu/libcudnn.so /usr/lib/x86_64-linux-gnu/libnvinfer.so -ldl -lcudadevrt -lcudart_static -lrt -lpthread -ldl 
cd /workspace/tensorRT/version-7.1/TensorRT/build/plugin && /usr/local/bin/cmake -E cmake_symlink_library ../out/libnvinfer_plugin.so.7.1.3 ../out/libnvinfer_plugin.so.7 ../out/libnvinfer_plugin.so
make[2]: Leaving directory '/workspace/tensorRT/version-7.1/TensorRT/build'
[100%] Built target nvinfer_plugin
make[1]: Leaving directory '/workspace/tensorRT/version-7.1/TensorRT/build'
/usr/local/bin/cmake -E cmake_progress_start /workspace/tensorRT/version-7.1/TensorRT/build/CMakeFiles 0

ledinhtri97 on 11 Nov 2020

Oh I see the issue - it looks like the exports map is hiding the symbols. Could you try replacing the contents of that file with:

/* Hides all symbols except those specified in the global section */
{
  global:
    createInferBuilder_INTERNAL;
    createInferRuntime_INTERNAL;
    createInferRefitter_INTERNAL;
    getInferLibVersion;
    createRPNROIPlugin;
    createNormalizePlugin;
    createPriorBoxPlugin;
    createAnchorGeneratorPlugin;
    createNMSPlugin;
    createLReLUPlugin;
    createReorgPlugin;
    createRegionPlugin;
    createClipPlugin;
    createBatchedNMSPlugin;
    createSplitPlugin;
    createInstanceNormalizationPlugin;
    getPluginRegistry;
    initLibNvInferPlugins;
    extern "C++" {
      nvinfer1::*;
      nvcaffeparser1::*;
      nvonnxparser::*;
      nvuffparser::*;
    };
  local: *;
};

pranavm-nvidia on 11 Nov 2020

👀1 ❤1

It seem worked, when have nvinfer1::plugin::*, but createBatchedNMSPlugin and other createPlugin still dont have, please check this again @pranavm-nvidia

ledinhtri97 on 12 Nov 2020

@ledinhtri97 That's expected since we don't expose this function in the OSS plugin header. You can use the plugin registry to achieve the same functionality, as mentioned here.

pranavm-nvidia on 12 Nov 2020

Some updates in https://github.com/NVIDIA/TensorRT/pull/952. @ledinhtri97 the create*Plugin methods are only supported in the plugin library shipped with GA builds and not the open source repo (since these are deprecated).