Problems compiling 3D-Caffe on Fedora24 about 3d-caffe HOT 4 CLOSED

On 13 Jan 2017, at 05:06, Jordina Torrents Barrena ***@***.***> wrote: Dear Fausto, thanks for your great work on 3D-Caffe! We were able to compile and run the standard Caffe.rc3 on Fedora 24 system. However, when we try to compile 3D caffe we stop with the following error: `Building NVCC (Device) object src/caffe/CMakeFiles/cuda_compile.dir/layers/cuda_compile_generated_absval_layer.cu.o Generating /home/jtorrents/CUDA/3D_Caffe/build-cmake/src/caffe/CMakeFiles/cuda_compile.dir/layers/./cuda_compile_generated_absval_layer.cu.o /usr/local/cuda-8.0/bin/nvcc /home/jtorrents/CUDA/3D_Caffe/src/caffe/layers/absval_layer.cu -c -o /home/jtorrents/CUDA/3D_Caffe/build-cmake/src/caffe/CMakeFiles/cuda_compile.dir/layers/./cuda_compile_generated_absval_layer.cu.o -ccbin /usr/lib64/ccache/cc -m64 -DUSE_CUDNN -DWITH_PYTHON_LAYER -DGTEST_USE_OWN_TR1_TUPLE -Xcompiler ,\"-std=c++98\",\"-fPIC\",\"-Wall\",\"-Wno-sign-compare\",\"-Wno-uninitialized\",\"-O2\",\"-DNDEBUG\" -gencode arch=compute_20,code=sm_20 -gencode arch=compute_20,code=sm_21 -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -Xcudafe --diag_suppress=cc_clobber_ignored -Xcudafe --diag_suppress=integer_sign_change -Xcudafe --diag_suppress=useless_using_declaration -Xcudafe --diag_suppress=set_but_not_used -Xcompiler -fPIC -DNVCC -I/usr/local/cuda-8.0/include -I/home/jtorrents/CUDA/3D_Caffe/src -I/usr/include -I/home/jtorrents/CUDA/3D_Caffe/build-cmake/include -I/usr/local/cuda-8.0/include -I/home/jtorrents/CUDA/CuDNN_v3/include -I/usr/include/opencv -I/usr/include/python2.7 -I/usr/lib64/python2.7/site-packages/numpy/core/include -I/home/jtorrents/CUDA/3D_Caffe/include -I/home/jtorrents/CUDA/3D_Caffe/build-cmake nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning). /home/jtorrents/CUDA/3D_Caffe/include/caffe/util/cudnn.hpp(141): error: too many arguments in function call /home/jtorrents/CUDA/3D_Caffe/include/caffe/util/cudnn.hpp(160): error: identifier "CUDNN_NOT_PROPAGATE_NAN" is undefined /home/jtorrents/CUDA/3D_Caffe/include/caffe/util/cudnn.hpp(160): error: too many arguments in function call /home/jtorrents/CUDA/3D_Caffe/include/caffe/util/cudnn.hpp(179): error: identifier "CUDNN_NOT_PROPAGATE_NAN" is undefined /home/jtorrents/CUDA/3D_Caffe/include/caffe/util/cudnn.hpp(179): error: argument of type "int" is incompatible with parameter of type "const int *" /home/jtorrents/CUDA/3D_Caffe/include/caffe/util/cudnn.hpp(179): error: too many arguments in function call 6 errors detected in the compilation of "/tmp/tmpxft_000074ab_00000000-5_absval_layer.cpp4.ii"` The configuration is as follows: -- ******************* Caffe Configuration Summary ******************* -- General: -- Version : <TODO> (Caffe doesn't declare its version in headers) -- Git : unknown -- System : Linux -- C++ compiler : /usr/lib64/ccache/c++ -- Release CXX flags : -O2 -DNDEBUG -std=c++98 -fPIC -Wall -Wno-sign-compare -Wno-uninitialized -- Debug CXX flags : -g -std=c++98 -fPIC -Wall -Wno-sign-compare -Wno-uninitialized -- Build type : Release -- -- BUILD_SHARED_LIBS : ON -- BUILD_python : ON -- BUILD_matlab : OFF -- BUILD_docs : ON -- CPU_ONLY : OFF -- -- Dependencies: -- BLAS : Yes (Open) -- Boost : Yes (ver. 1.60) -- glog : Yes -- gflags : Yes -- protobuf : Yes (ver. 2.6.1) -- lmdb : Yes (ver. 0.9.18) -- Snappy : Yes (ver. 1.1.3) -- LevelDB : Yes (ver. 1.12) -- OpenCV : Yes (ver. 2.4.12.2) -- CUDA : Yes (ver. 8.0) -- -- NVIDIA CUDA: -- Target GPU(s) : Auto -- GPU arch(s) : sm_20 sm_21 sm_30 sm_35 sm_50 -- cuDNN : Yes (v3) -- -- Python: -- Interpreter : /usr/bin/python2.7 (ver. 2.7.12) -- Libraries : /usr/lib64/libpython2.7.so (ver 2.7.12) -- NumPy : /usr/lib64/python2.7/site-packages/numpy/core/include (ver 1.11.0) -- -- Documentaion: -- Doxygen : No -- config_file : -- -- Install: -- Install path : /home/jtorrents/CUDA/3D_Caffe/build-cmake/install -- -- Configuring done Thanks and regards! Jordina — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

On 13 Jan 2017, at 09:17, Jordina Torrents Barrena ***@***.***> wrote: Hi Fausto! Thank you very much for your answer :) I solved the previous error through the CuDNN v5 installation (instead of v3). Although I perfectly compiled your 3D-Caffe code, I cannot run all test. For instance, some errors like that appear: 8 tests from CuDNNNeuronLayerTest/0, where TypeParam = float [ RUN ] CuDNNNeuronLayerTest/0.TestSigmoidCuDNN F0113 15:10:48.077026 7644 cudnn_sigmoid_layer.cu:16] Check failed: status == CUDNN_STATUS_SUCCESS (3 vs. 0) CUDNN_STATUS_BAD_PARAM *** Check failure stack trace: *** @ 0x7f6776d2a060 (unknown) @ 0x7f6776d2be60 (unknown) @ 0x7f6776d29c13 (unknown) @ 0x7f6776d2c8a9 (unknown) @ 0x7f6777df18f6 caffe::CuDNNSigmoidLayer<>::Forward_gpu() @ 0x77ca3a caffe::Layer<>::Forward() @ 0x9c9476 caffe::CuDNNNeuronLayerTest_TestSigmoidCuDNN_Test<>::TestBody() @ 0xba4694 testing::internal::HandleExceptionsInMethodIfSupported<>() @ 0xb9e01a testing::Test::Run() @ 0xb9e168 testing::TestInfo::Run() @ 0xb9e245 testing::TestCase::Run() @ 0xb9f467 testing::internal::UnitTestImpl::RunAllTests() @ 0xb9f763 testing::UnitTest::Run() @ 0x772b39 main @ 0x7f67658e1731 __libc_start_main @ 0x778bf9 _start /bin/sh: line 1: 7644 Aborted (core dumped) /home/jtorrents/CUDA/3D_Caffe/build-cmake/test/test.testbin --gtest_shuffle Do you know why this happens?... I believe it is related to the CuDNN version, but now I have installed v5 with CUDA 8.0 (my GPU is the following: GeForece GTX 950M). Best regards, Jordina — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.