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.
233
questions
0
votes
0
answers
12
views
ffmpeg: filter frame using cuda
Is it possible to filter frames on nvidia GPU side?
For intel qsv it looks like
[sprites_src]vpp_qsv=framerate=1/10,scale_qsv=-1:160:mode=$mode:format=rgb32,hwdownload,format=rgb32[sprites_out];
For ...
- 160
2
votes
1
answer
66
views
Is it possible to mix CUDA compute capabilities with two different gpus on a single system?
OS is Debian 10. I currently have a Geforce 1060 GTX that I'm trying to use for various CUDA tasks. However, there is no way to "renice" a process that runs on a video card, so when my tasks ...
- 307
1
vote
1
answer
217
views
How do I get CUDA 11.7 on Arch Linux?
This should be pretty straightforward, pacman search for it, if not, yay search for it, if not, try to find some other AUR for it. If none of that works, use the official binary blob from Nvidia.
...
- 745
0
votes
1
answer
87
views
Does any gaming laptop with NVIDIA GPU support CUDA?
Does any gaming laptop with NVIDIA GPU support CUDA?
Or, do I have to be careful about purchasing one?
- 111
1
vote
1
answer
454
views
FFmpeg & CUDA - scale video and overlay image
I want to scale input video to 1920x1080, then overlay image
And I use next command:
ffmpeg -i /video1.mp4 -i /overlay.png \
-filter_complex "[0:v][1:v] overlay=100:820:enable='between(t,...
- 21
0
votes
0
answers
21
views
How does Windows monitors/calculates GPU utilization?
I've been coding in CUDA recently, and I launched 2^31 kernels onto my GPU to test out. The program took around 63 seconds to complete, but my GPU utilization (seen in Windows Task Manager) didn't get ...
5
votes
2
answers
2k
views
Cuda in WSL2: "GPU access blocked by the operating system"
I am attempting to install cuda in WSL2 for use with TensorFlow (as per the TF instructions, which include the NVIDIA instructions).
The Windows-side tool nvidia-smi.exe works and gives the following ...
- 151
0
votes
1
answer
2k
views
Unable to get CUDA to work on conda enviornment Ubuntu 22.04
I am trying to install Cuda 11.3 (since i understand this is the most stable one and any version after leads to compatibility issues). I have Ubuntu 22.04 LTS and I followed these instructions: https:/...
- 11
0
votes
0
answers
118
views
How to install cuda and nvidia driver for Tesla C2050?
I installed win 20 and CentOS 7 system in my computer, and then I tried to install cuda for Tesla C2050 GPU in CentOS 7 but cannot success. For installing cuda I ran the following steps
(1)# lspci | ...
- 101
1
vote
0
answers
492
views
WLS 2 on Windows 11 (Ubuntu 22) freezes after ~15 hours of running
This appears to be a similar problem to this unanswered question.
Usually I have a tail -f nohup.out command running in one pane, and nvidia-smi running in the other pane (refreshing every second) to ...
- 11
0
votes
0
answers
232
views
CUDA Toolkit Minimum Required Driver Version For Development and Docker Deployment
I know binary compatibility only need R450 Driver for CUDA 11.x
https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html
The documentation says: "Running a CUDA application requires the ...
- 101
0
votes
1
answer
131
views
How to parallelize a Python simulation script on a GPU with CUDA?
I want to test run some finite-difference time domain (FDTD) simulations using parallelization on a GPU and compare the speed to CPU-based simulation runs. This is basically my very first attempt at ...
- 227
0
votes
0
answers
54
views
Using new CUDA for debuging on new GPU and old Nvidia GPU for displys
I need to debug a CUDA kernel. I know, that I need two GPU's for it. And I have them: Titan Black (cc 35) and GTX 460 (cc 21). I work in Manjaro linux and I can install old drivers from AUR. Nvidia ...
0
votes
1
answer
168
views
Can I delete directories in /var without verified signatures (NO_PUBKEY Error) in Ubuntu 20.04?
When I execute sudo apt-get update command I get the following terminal output:
Get:1 file:/var/cuda-repo-ubuntu2004-11-3-local InRelease
Ign:1 file:/var/cuda-repo-ubuntu2004-11-3-local InRelease
...
- 1
1
vote
0
answers
45
views
Is it possible to use cuda or opengl in a container without passthrough the physical gpu?
After googling how to run cuda in a container, I am confused why most articles requires "passthrough" the gpu to the container: the container does not have own kernel, how does the kernel ...
- 133
0
votes
0
answers
230
views
Manually editing cuda version loaded by Tensorflow
Each version of tensorflow is only compatible with one version of cuda and should automatically look for and try to load the correct library version. For some reason, my tensorflow installation is ...
2
votes
1
answer
3k
views
Accurately extract frames every N seconds using FFmpeg
The Usage
I extract images from videos using ffmpeg.
I dump one down-scaled image every 10 seconds, inclusive, that I combine into montages with imagemagick. These montages again are used to show a ...
- 123
1
vote
1
answer
622
views
FFmpeg: How can I hstack two videos, and scale just one up with scale_cuda?
This is a command I used that worked with scale_npp instead of scale_cuda
ffmpeg -c:v h264_cuvid -i "Input1.mkv" -c:v h264_cuvid -i "Input2.mkv" -c:v h264_nvenc -filter_complex &...
- 173
1
vote
1
answer
3k
views
Cuda libraries not found for cuda 11 for ubuntu 20.04
I am using Nvidia VM on Azure with Ubuntu 20.04 I already had nvidia and cuda installed but hile running my program it still showed libraries not found
nvidia-smi
+-------------------------------------...
- 111
0
votes
1
answer
764
views
How to configure nvidia-cuda-mps-control so that there can be multiple users running on the same system
I have, for example, a 4 GPU system. I would like to be able to have 2 users each using MPS on 2 GPUs. In reading the nvidia documentation it looks like I would set CUDA_VISIBLE_DEVICES to specify ...
1
vote
1
answer
776
views
How to check the health of a GPU
I have a GTX 1080Ti which I have been using as my personal deep learning card for about 3,5 years. Lately, I'm getting a lot of obscure CUDA errors. Training and inference code that was running fine ...
- 111
0
votes
0
answers
428
views
Tensorflow GPU not working on Ubuntu 20.04 (error message)
So, I tried to go through several tutorials (the latest being this one) to set up GPU support for TensorFlow on Ubuntu 20.04 and I always get the following message:
joris@joris-N751JK:~$ python3.9
...
- 101
0
votes
0
answers
142
views
TensorFlow GPU on Ubuntu 20.04
The Tensorflow documentation details how to install Tensorflow with GPU support for Ubuntu 16.04 and 18.04. I can't find anything up to date for Ubuntu 20.04 however.
How do you install Tensorflow ...
- 101
0
votes
1
answer
33
views
Running DL Training on CUDA GPU (util 98%) does not impact gaming? Are CUDA cores not Render Cores?
I downloaded CUDA 11 to run some heavy (and I mean it, heavy) DL model training tasks that use low memory (2GB). I'm rocking a 3070 Ti with 8GB. However in task manager I can see CUDA utilization is ...
- 101
2
votes
0
answers
197
views
Starting previously running Docker containers with GPU support
I recently set up WSL2 (Windows 10, 21H2) with CUDA/Docker support to train some neural networks. For this I followed the instructions from this guide: https://docs.nvidia.com/cuda/wsl-user-guide/...
- 131
-1
votes
1
answer
60
views
Is there anyway to decode a mp4 to bmp images with cuda hwaccel
ffmpeg 3.4.9
ubuntu16.04
I use follow command to decode a mp4 to bmp files (all in one),then I split 1.bmps to image array.
ffmpeg -y -i 1632035082_1.mp4 -r 20/1 -vcodec bmp -f image2pipe - > ...
- 1
1
vote
1
answer
2k
views
"Cannot load libcuda.so.1" when running ffmpeg with hardware acceleration on WSL2 Windows
I am following NVIDIA's doc for building ffmpeg with hardware acceleration capabilities. I have also installed the appropriate driver for my GeForce GPU and did the CUDA toolkit setup for WSL2. I am ...
- 111
1
vote
0
answers
287
views
try compile ffmpeg with cuda 11.1 on window10 but get error when run make -j 8
I am trying to compile ffmpeg with cuda 11.1 on Windows. I run configure like this:
./configure --arch=x86_64 --target-os=mingw64 --enable-cuda --enable-cuvid --enable-nvenc --enable-nonfree --enable-...
- 11
1
vote
0
answers
506
views
cv2.cuda.getCudaEnabledDeviceCount() return 0 after visual studio build and with_cuda is set true for opencv using windows10
I want gpu to work in opencv , after CMake with_cuda true and build of (ALL_Build, Install) with no errors
I found GPU not detected
I use windows 10 I tried 3 times to build opencv , but Gpu not work ...
- 111
3
votes
0
answers
3k
views
WSL 2 cannot see NVIDIA GPU drivers
I have Windows 10 along with Ubuntu 20.04.01 GNU/Linux 5.10.16.3-microsoft-standard-WSL2 x86_64 installed (latest for WSL2 - please see edits below- this has changed). I'm a Windows Insider user and ...
- 81
0
votes
1
answer
293
views
13.022698] nvidia-modeset: Version mismatch: nvidia.ko(430.64) nvidia-modeset.ko(460.32.03)
I have been trying to get cuda 10.2 installed on ubuntu 18.04. I am getting this mismatch error. There are similar posts on stackoverflow. I tried deleting dpkg lines, and repeat the install script, ...
- 101
0
votes
2
answers
4k
views
CUDA support for RTX 3080 (Laptop GeForce Notebook Products)
Does RTX 3080 for laptops already support CUDA and cuDNN for Deep learning requirements? I want to buy a new Razor Blade 15 with RTX 3080 and I don’t see currently any listing of this GPU on Nvidia ...
0
votes
1
answer
2k
views
Failled CUDA installation Ubunutu 18.04.5
I'm trying to install the CUDA toolkit with Ubuntu 18.04.5 for my GTX 3060 Ti. I've downloaded the file cuda_11.2.1_460.32.03_linux.run from NVidia's site followed the instructions and when then there ...
- 1
3
votes
1
answer
229
views
Is there a way to use Linux on Windows while utilizing nvidia CUDA?
VirtualBox cannot use CUDA.
Is there a way around without using dual-boot?
0
votes
0
answers
1k
views
ffmpeg not setting up Nvidia card if it's not the first adapter
I just set up a 'new' PC with built-in Intel video and an Nvida card: mostly to speed up video processing with ffmpeg and other programs. At first the built-in Intel was disabled, running only the ...
0
votes
1
answer
644
views
Error after installing CUDA on WSL 2 - RuntimeError: No CUDA GPUs are available
I am trying to install CUDA on WSL 2 for running a project that uses TorchAudio and PyTorch. I have done the steps exactly according to the documentation here.
However, when I run my required code, I ...
- 101
3
votes
0
answers
4k
views
NVIDIA: “RuntimeError: No CUDA GPUs are available”
I am implementing a simple algorithm with PyTorch on Ubuntu. Two times already my NVIDIA drivers got somehow corrupted, such that running an algorithm produces this traceback:
Traceback (most recent ...
- 31
3
votes
1
answer
5k
views
Set keyframe interval in ffmpeg when using nvenc
OBS uses nvenc when encoding live streams. In the guide "NVIDIA NVENC OBS Guide", an advanced option is listed as:
Keyframe Interval: Set to 2. Streaming platforms may limit what you can ...
- 193
4
votes
2
answers
20k
views
Understanding pixel format and profile when encoding 10-bit video in ffmpeg with nvenc
I am trying to encode a 10-bit H.265 video from a 8-bit H.264 source using ffmpeg with CUDA hardware acceleration.
Without hardware acceleration, a typical command would be ffmpeg -i input.mkv -...
- 193
1
vote
2
answers
690
views
Power of NVIDIA GPU out of limits (Nvidia A100)
I am executing a CUDA kernel in my A100 GPU and I've realized that the power consumption at some points is higher than nvidia-smi given range:
The picture has been taken from nvtop.
Is it something ...
- 151
1
vote
1
answer
1k
views
Cannot open NVIDIA Visual Profiler on Ubuntu 20.04
I've installed CUDA kit via apt-get following CUDA installation documentation. However, when I attempt to open NVIDIA Visual Profiler, I can't open it. By the log file I suppose it has something to do ...
- 131
1
vote
0
answers
886
views
NVIDIA-SMI has failed because it couldn’t communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running
System Spec:
OS: ubuntu 18.04 GPU:
Quadro RTX 4000
I have a fresh installed machine with ubuntu 18.04. To install CUDA 11 toolkit these are the command that i ran :
wget https://developer.download....
- 11
0
votes
1
answer
189
views
Does Nvidia rtx series still features plain cuda cores or is everything replaced by Tensor cores?
For a long time, consumer Nvidia Graphic cards supportted having plain cuda cores despite not being branded for accelerators.
The problem in my case is while Tensor cores allow to perform cuda, they ...
- 1,780
1
vote
0
answers
223
views
Cuda enabled FFMPEG not showing any GPU usage
I've compiled ffmpeg 4.3.1 with cuda support using the Toolkit 11. However, when I run ffmpeg against an rtsp stream and transcode into mp4, the process never shows any GPU usage in ProcessExplorer.If ...
- 11
4
votes
1
answer
6k
views
How to install pytorch FROM SOURCE (with cuda enabled for a deprecated CUDA cc 3.5 of an old gpu) using anaconda prompt on Windows 10?
I have (with the help of the deviceQuery executable in C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vX.Y\extras\demo_suite
according to https://forums.developer.nvidia.com/t/what-is-the-compute-...
- 1,715
0
votes
0
answers
857
views
How do you get FFmpeg script -hwaccel overlay_cuda to fade in and out an image?
The following FFmpeg hardware-acceleration script works fine. But when I try to apply a fade in and out to the overlayed image it breaks. Is there a solution?
Working:
ffmpeg -hwaccel cuda -...
1
vote
1
answer
282
views
Installing CUDA in GCP VM instance ( Debian 10.x)
I am just coming up to speed with GPU & CUDA and going around in circles.
My GPU is Nvidia Tesla T4. However, Google documentation is lacking
https://cloud.google.com/compute/docs/gpus/install-...
- 111
0
votes
0
answers
129
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
2k
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
1k
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