Questions tagged [cuda]
Computing architecture developed by Nvidia that allows programs to process algorithms in the graphics card GPU, instead of the computer's CPU.
236
questions
0
votes
0
answers
158
views
Laptop won’t boot after installing CUDA tookit
Today I installed the CUDA toolkit on my laptop (NVIDIA GeForce GTX 1070 graphics card).
After a restart my laptop won’t boot anymore.
I, nor the CUDA installer, created a System Restore Point. I ...
- 123
0
votes
1
answer
3k
views
how to uninstall and reinstall nvidia-drivers (linux 20.04)
I have recently installed ubuntu 20.04 on my laptop and facing some issues running tensorflow1.X-gpu version . The problem is with the installation of cuda, cudnn and nvidia-drivers . I want to use ...
0
votes
1
answer
2k
views
What exactly are compute shaders? [closed]
What are compute shaders?
Refined Question:
I need to be able to write a shader knowing whether it will be processed GPU side or if I need to do it CPU side. For instance I would like to create tiles ...
- 13
2
votes
0
answers
1k
views
Python: Do I need to install Tensorflow and Pytorch in two different conda environments?
Intro:
I do not know whether I should install pytorch and tensorflow in separated environments if I want to install both.
For conda tensorflow, I find at https://machinelearningspace.com/installing-...
- 2,331
0
votes
1
answer
2k
views
CUDA driver compatibility with nvfortran compiler from nvidia hpc sdk 20.7
I have installed the nvidia compilers (nvcc, nvfortran). I downloaded it from the NVIDIA HPC SDK v20.7 site. I then compiled non-cuda codes with nvfortran, and it works fine. But when i compile the ...
0
votes
1
answer
960
views
Installing CUDA Toolkit 11 when your computer CUDA version says 10.1
I suspect that CUDA is not installed in my windows 10 machine, so I am planning to download and install it. For that I would use this link that allows me to get CUDA Toolkit 11.
However if I do nvidia-...
- 856
3
votes
1
answer
15k
views
Do I need to install cuda separately after installing the NVIDIA display driver?
Is cuda already included in the installation of the graphics card display driver?
The card has 192 cuda cores according to https://www.asus.com/Graphics-Cards/GT710-SL-2GD5/specifications/, and Cuda 3....
- 2,331
1
vote
0
answers
325
views
Command 'nvcc' not callable from shell script
For my shell script, it is critical to check whether CUDA is installed by running nvcc --version. If I try to invoke it, it exits with following error:
./cuda.sh: line 17: nvcc: command not found
...
- 111
1
vote
0
answers
329
views
All the folders under CUDA folder are Empty. (After installed the CUDA 10.1 on Ubuntu 18.04)
I met strange a problem after I installed the CUDA 10.1 on my Ubuntu 18.04 server. I found all the files under the CUDA folder are Empty! Can anyone help me with this issue? Thank you in advance for ...
2
votes
0
answers
432
views
Why peer-to-peer (P2P) access between two Tesla K40c GPUs fails in CUDA?
I would like to run a CUDA C program using two Tesla K40 devices and enable peer-to-peer (P2P) between them as my data will be shared among the devices. My computer has the following deviceQuery ...
0
votes
1
answer
189
views
Is there a way to get Cuda support for an older GeForce GT?
An older GPU doesn't seem to support Cuda? I am under the understanding that Cuda is only software so I assume it could somehow be forced to use it.
It's an GeForce GT 530
PS: I've been told that my ...
0
votes
1
answer
28
views
Failure in running Faster R-CNN repo
I have a MacBook Pro 16in with the following GPU: AMD Radeon Pro 5300M.
I have been trying to run the demo in the Faster R-CNN repo here but I am not being able to run the demo.
I think it might be ...
- 101
0
votes
2
answers
3k
views
“Pixel format 'yuvj420p' is not supported” when converting image to video using CUDA hardware acceleration
I'm trying to convert an image to a video using CUDA hardware acceleration but I can't get it to work.
The basic command I have is this one:
ffmpeg -y -hwaccel cuvid -i tests/MediaSamples/...
- 6,119
2
votes
1
answer
2k
views
How do I install Pytorch 1.3.1 with CUDA enabled
I have a conda environment on my Ubuntu 16.04 system.
When I install Pytorch using:
conda install pytorch
and I try and run the script I need, I get the error message:
raise AssertionError("Torch ...
- 185
0
votes
0
answers
442
views
How to use GPU acceleration on Docker without DirectX on Windows Containers
Microsoft documents that GPU acceleration can be enabled on frameworks built on top of DirectX here
NVidia's base Docker image for CUDA is Linux based here
Question: How to enable GPU acceleration ...
- 1
2
votes
1
answer
40
views
cuda v100 and RTX2080 bandwidth fluctuations
I have a simple cuda kernel (adding two vectors of size N) pretty similar to to this cuda blog here. I only changed a few things, e.g. running the measurement over various sample. So, let this run for,...
- 31
1
vote
0
answers
108
views
When I am installing Nvidia driver and Cuda, CentOS 7 cannot boot [image]
What is the reason preventing CentOS7 to boot when installing Nvidia driver and Cuda.
I have tried two times. The first time only installed the driver and then when I restart the computer it hangs on ...
- 111
3
votes
1
answer
2k
views
How to kill all processes using a given GPU?
I use the CUDA toolkit to perform some computations on my Nvidia GPUs on Windows 7 SP1 x64 Ultimate. How to kill all processes that use a given GPU? (killing all processes at once)
- 21.3k
1
vote
2
answers
2k
views
Installing CUDA 9.2 with VS 2019
I am trying to install CUDA 10.1 and 9.2 on my Windows 10 x64 system. (I need both because some software still relies on 9.2 to build and run.) I have Visual Studio 2019 installed. Installing 10.1 (...
- 1,307
3
votes
2
answers
12k
views
Is there a way to unite WSL2 Ubuntu, Tensorflow-GPU, CUDA, and Pycharm[windows]?
I have a WSL 2 Ubuntu with Anaconda 3 installed in Ubuntu WSL 2.
Then I installed PyCharm in Windows 10.
Then I created an environment for tensorflow-gpu in WSL2.
I installed CUDA and cuDNN in Windows ...
1
vote
0
answers
233
views
ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory
I am using ubunutu 18.04
this is the result of nvidia-smi:
Mon Oct 28 12:00:58 2019
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 435.21 ...
- 123
0
votes
1
answer
7k
views
InternalError: cudaGetDevice() failed. Status: cudaGetErrorString symbol not found
I run this code:
tf.test.is_gpu_available( cuda_only=False, min_cuda_compute_capability=None )
I get the following error:
2019-10-25 18:25:20.855191: I
tensorflow/core/platform/...
0
votes
1
answer
779
views
How to fix the CUDA out of memory error on EC2 instance while running inference pipeline?
I am trying to run a complete inference pipeline which takes a number of images and first runs Object Detection (multiprocessing) and further runs a Classifier (multiprocessing). I have tested the ...
- 101
0
votes
0
answers
1k
views
NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. (even after driver reinstall )
i had faced with an issue on nvidia driver for 2080Ti (need to install nvidia driver + CUDA for torch usage).
me@server:~$ python3
Python 3.6.8 (default, Aug 20 2019, 17:12:48)
[GCC 8.3.0] on linux
...
0
votes
1
answer
126
views
CUDA documentation does not mention compatibility requirements for Debian systems
I often need to recall the compatibility between host and device compilers and refer to the documentation page
https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#system-requirements ...
- 435
0
votes
1
answer
382
views
what is the limitation on max cuda instances for video transcoding?
I have set up Nvidia V100 card on Azure Nc6s_v3 virtual machine. I have set up the latest CUDA driver and SDK. I compiled sample programs from "Video_Codec_SDK_9.0.20" and verified the max parallel ...
- 111
1
vote
0
answers
1k
views
RuntimeError: cuda runtime error (59) : device-side assert triggered a
I checked other suggestions, but other people used RNN Networks and data labels.
In my case, everything was working fine yesterday, but suddenly my code is not working anymore.
I am trying to run ...
- 111
0
votes
0
answers
424
views
My graphics card does not have CUDA 9.2, what is this?
I'm trying to run a graphics editing program on my PC. It tells me I need to have CUDA 9.2 for the graphics card - I'm running an NVIDIA GeForce 210, although I may update this for video editing on ...
- 1
8
votes
1
answer
3k
views
Pycharm debugger does not work with pytorch and deep learning
This question is somewhere in between Stackoverflow and Superuser - in my opinion at least, so feel free to point me to SO if this is the wrong place (in your opinion ;)).
Problem:
If I place a ...
- 183
0
votes
1
answer
385
views
CUDA 10.1 & Mac OS X 10.13.6 Driver Version Insufficient
I installed all the CUDA packages, drivers, samples, and cuDNN.
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2019 NVIDIA Corporation
Built on Sun_Jul_28_19:14:47_PDT_2019
Cuda compilation ...
1
vote
1
answer
3k
views
How to remove the previous cuda 7.5 and install new version of v. 10 in centos 7?
There is a process of the installation of cuda 10.0 How to install NVIDIA CUDA Toolkit on CentOS 7 Linux where a repository is installed through, and then cuda can be installed by
rpm -i cuda-repo-...
- 121
2
votes
0
answers
600
views
Fail message setting up docker container
I am trying to set up a docker container on my linux 18.04 system with an nvidia/cuda image but i cant make it run.
When i enter
sudo nvidia-docker run --rm nvidia/cuda nvidia-smi
i get this error ...
- 21
0
votes
0
answers
126
views
Nsight eclipse is pointing to wrong nvcc path in Ubuntu 16.04
I have successfully installed nsight eclipse through cuda toolkit 10.1. I am able to compile sample cuda programs through command line but facing the error with nvcc path. Eclipse is pointing wrongly ...
1
vote
0
answers
213
views
Trouble running Tensorflow-gpu GPT-2 model
I'm trying to run OpenAI's GPT-2 model using the Tensorflow GPU running on my NVidia SLI GTX-780 though I keep getting errors regardless of what versions I run or what I do.
File "C:\Users\myUser\...
- 11
1
vote
1
answer
2k
views
Not able to uninstall CUDA 10.1 on Ubuntu 16.04.6 LTS
I am trying to uninstall Cuda 10.1 on Ubuntu 16.04.6 LTS but I am not able to uninstall it using the instructions given on nvidia website.
I have installed CUDA 10.1 on Ubuntu 16.04.6 LTS using the ...
3
votes
1
answer
1k
views
Setting NVreg_RestrictProfilingToAdminUsers=0 on Centos 7
I've run both nsight profiler and nsight compute and i get the same issue:
nvprof log: /home/centos/nvvp_workspace/.metadata/.plugins/com.nvidia.viper/launch/0/nvprof_21332.log
==21332== Warning: ...
- 131
2
votes
1
answer
497
views
Can't seem to get driver working in Cuda 10.0 Installation
Intro
Hi, I was recently given access to a remote machine with a GPU that I can use to accelerate my research. However, I am having a lot of trouble getting it set up. I went through the Nvidia Linux ...
- 21
1
vote
0
answers
187
views
Installation of CUDA 9.0 in windows 10
I am trying to install Cuda 9.0 in windows 10. for the purpose of running this example:
https://github.com/andi611/Apriori-and-Eclat-Frequent-Itemset-Mining
Python Version: 3.6.5 ( Anaconda)
...
- 111
0
votes
1
answer
2k
views
CUDA version mismatch on Ubuntu 18.04
I'm trying to use the CUDA toolkit on Ubuntu 18.04.
I installed the CUDA toolkit by using
sudo apt install nvidia-cuda-toolkit gcc-6
This installed CUDA 9.1, verified by using nvcc --version, ...
- 1
1
vote
1
answer
326
views
Looking for a way to get PyCUDA with default MacOS NVIDIA drivers
I have a CUDA NVIDIA card ( GeForce GT 750 M) and I would like to use the default NVIDIA drivers of MacOS to be able to install PyCUDA like I did with NVIDIA drivers.
Is it possible ? For the moment, ...
user342714
1
vote
1
answer
723
views
Anyone managed to run multiple NVidia RTX 2080Ti cards in parallel with NCCL (caffe, tensorflow) without NVLink bridge?
I am trying to run DNN training on machine with 8x GPU cards. However it appears that RTX 2080Ti cards (and may be Titan RTX) can not be run in parallel without NVLink bridge. Which means you can only ...
52
votes
6
answers
54k
views
Is it safe to keep the GPU on 100% utilization for a very long time?
I am currently performing number-crunching using CUDA on my GPU, an NVIDIA GeForce GTX 1050 Ti. These operations often take months to complete, and during that time I leave my PC on 24/7.
Is doing so ...
- 827
1
vote
1
answer
4k
views
Tensorflow-gpu: CUDA_ERROR_OUT_OF_MEMORY [closed]
I am relatively new to tensorflow and tried to install tensorflow-gpu on a Thinkpad P1 (Nvidia Quadro P2000) running with Pop!_OS 18.10. I installed tensorflow-gpu into a new conda environment and ...
- 11
0
votes
1
answer
597
views
Failed to initialize GPU device #0: unknown error
Each time I run the fit method for keras model, I got "...loss: nan - val_loss: nan"
System configuration:
windows 10
tensorflow 1.13.1
Keras 2.2.4
GPU settings:
Type: ...
0
votes
1
answer
10k
views
Test tensorflow-gpu failed with Status: CUDA driver version is insufficient for CUDA runtime version (which is not true)
I have the following configuration:
SUSE Linux Enterprise Server 12 SP3 (x86_64)
CUDA Toolkit: CUDA 9.2 (9.2.148 Update 1)
CUDA Driver Version: 396.37
According to NVIDIA just right (https://docs....
0
votes
1
answer
453
views
install nvidia-driver418 and cuda9.2.-->CUDA driver version is insufficient for CUDA runtime version
everyone,
My machine is ubuntu16.04
I got 2 Tesla P100 ,
Firstly,I install nvidia-driver 418, the result is good.
Then Then I install cuda9.2(The reason why I didn't choose cuda10.0 is that I want to ...
- 1
1
vote
2
answers
2k
views
Encountered Bus Error while trying to run code in CUDA
This was the code that I was trying to run on my computer when I encountered this error.
./l
[Vector addition of 10 elements]
Bus error (core dumped)
I tried this
>>>> nvcc -o l ...
- 125
3
votes
1
answer
1k
views
internal error during nvprof profiling
I am using the nvprof to collect some information about applications. I can run "nvprof a.out" successfully(assuming we have a CUDA program called a.out), but I can not run "nvprof --metrics all a.out"...
- 31
2
votes
2
answers
2k
views
C++ Compiler from Visual Studio 2017 and NVidia's CUDA?
I am attempting to set up PyTorch to work my laptop's on board GTX 1050 GPU. Following this I have attempt to set up CUDA, I have been following the guide set up by Nvidia here. As far as I can tell I ...
- 131
4
votes
0
answers
465
views
Tensorflow gpu was working but is not working anymore error- Failed to get convolution algorithm. This is probably because cuDNN failed to initialize
So I have cuda 10 installed along with tensorflow gpu 1.13.1 and everything was working just fine until I imported a VGG16 model. After running the model once and everything working just fine, when ...
- 41