![]() #$ NVVMIR_LIBRARY_DIR=/usr/local/cuda-9.0/bin/./nvvm/libdevice $ /usr/local/cuda-9.0/bin/nvcc -verbose -ccbin g++ -I././common/inc -m64 -gencode arch=compute_30,code=sm_30 -o deviceQuery.o -c deviceQuery.cpp The output is $ cd /home/Kong/NVIDIA_CUDA-9.0_Samples/1_Utilities/deviceQuery/ I have no idea what’s wrong with it, please help me fix this problem. home/Kong/anaconda2/bin:/usr/local/cuda-9.0/bin:/home/Kong/anaconda2/bin:/usr/local/cuda-9.0/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. Here is the device information $ nvcc -versionĬopyright (c) 2005-2017 NVIDIA CorporationĬuda compilation tools, release 9.0, V9.0.176 ![]() It seems all the samples I compile have the problem with helper_cuda.h. Makefile:273: recipe for target 'deviceQuery.o' failed IDevice Co-op Kernel Launch: %s\n", operative ![]() /common/inc/helper_cuda.h:293:14: error: 'cudaErrorCooperativeLaunchTooLarge' was not declared in this scopeĬase cudaErrorCooperativeLaunchTooLarge :ĭeviceQuery.cpp: In function 'int main(int, char**)':ĭeviceQuery.cpp:172:84: error: 'struct cudaDeviceProp' has no member named 'cooperativeLaunch'Įrative Kernel Launch: %s\n", operativeĭeviceQuery.cpp:173:84: error: 'struct cudaDeviceProp' has no member named 'cooperativeMultiDeviceLaunch' /common/inc/helper_cuda.h:289:14: error: 'cudaErrorJitCompilerNotFound' was not declared in this scope /common/inc/helper_cuda.h:285:14: error: 'cudaErrorNvlinkUncorrectable' was not declared in this scope /common/inc/helper_cuda.h: In function 'const char* _cudaGetErrorEnum(cudaError_t)': In file included from deviceQuery.cpp:20:0: I am using Nvidia driver version 384.183, which satisfies the requirement for cuda-9.0 (version >=384).īut I can’t compile any samples, here is the example when I compile the samples in /home/Kong/NVIDIA_CUDA-9.0_Samples/1_Utilities/deviceQuery/, it comes out the errors: "/usr/local/cuda-9.0"/bin/nvcc -ccbin g++ -I././common/inc -m64 -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_70,code=compute_70 -o deviceQuery.o -c deviceQuery.cpp (Please see output of nvcc -version and nvidia-smi below). ![]() sudo apt-get install libcupti-devĮdit file sudo gedit /etc/ld.so.conf.d/cuda-9-0.I follow the official documentation carefully and It seems that I have successfully installed cudatoolkit 9.0. This library provides advanced profiling support. Step 10: The libcupti-dev library, which is the NVIDIA CUDA Profile Tools Interface. # Use OpenCV and other custom-built libraries.Įxport LD_LIBRARY_PATH=/usr/local/lib/:$LD_LIBRARY_PATH Post Installation step (path setting append in the file) gedit ~/.bashrc #(Append below lines without dashes)Įxport LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH sudo dpkg -i cuda-repo-ubuntu-local-cublas-performance-update-3_1.0-1_bħ. sudo dpkg -i cuda-repo-ubuntu-local-cublas-performance-update-2_1.0-1_bĦ. sudo dpkg -i cuda-repo-ubuntu-local-cublas-performance-update_1.0-1_bĥ. Step 9: Follow CUDA installation steps and install CUDA 9.0 Select appropriate settings for your machine as shown and in installer type select deb (local)ĭownload All Base Installer and Patch’s setup as shown Step 8: Download CUDA 9.0 Toolkit latest from check in Legacy Release if its old ( ) Step 7: Create a file and paste the following lines sudo gedit /etc/modprobe.d/nf (Paste Following without dash line and save changes) Step 6: Install latest nvidia drivers from package manager (use synaptic to see version) sudo apt-get install nvidia-384 Step 5: Update Repository list sudo apt-get update Step 4: Add Repository of Graphics Drivers sudo add-apt-repository ppa:graphics-drivers/ppa Sudo apt-get -purge -y remove 'libcupti* sudo dpkg -l | grep cuda- | awk '' | xargs -n1 sudo dpkg -purge Step 3: Remove old nvidia drivers and cuda setup sudo apt autoremove cuda sudo apt-get -purge -y remove 'cuda*' Step 2: Install Linux Headers (for installing aptitude “sudo apt install aptitude”) sudo aptitude -r install linux-headers-$(uname -r) Step 1: Update and upgrade your system sudo apt-get update & sudo apt-get upgrade -y
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |