Super User
Super User

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.

CUDA (Compute Unified Device Architecture) is a computing architecture developed by Nvidia, that allows programs to process algorithms in the graphics card GPU, instead of the computer's CPU.

Programs that make use of CUDA specially design some of their algorithms to make use of the GPU through Nvidia's framework. The program can then run both in the CPU and the GPU. Algorithms making use of CUDA will be sent to the GPU for processing, while algorithms that do not will be normally processed by the CPU.

CUDA is already being used by many systems. Particularly those that can take advantage of a GPU superior ability to perform floating-point calculations. These include scientific simulations and data processing, Distributed Computing and some encryption/decryption algorithms.

248 questions
398
votes
15 answers

Why are we still using CPUs instead of GPUs?

It seems to me that these days lots of calculations are done on the GPU. Obviously graphics are done there, but using CUDA and the like, AI, hashing algorithms (think bitcoins) and others are also done on the GPU. Why can't we just get rid of the…
ell
  • 4,054
53
votes
6 answers

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 safe? Am I risking a potential overheating of my…
Klangen
  • 847
48
votes
1 answer

How can I test my GPU memory/RAM?

I run MemTest86 a lot at work on customer's machines, and it's great for troubleshooting memory issues. My question is, how can I test that a GPU is starting to go? I know of programs like 3DMark to push the graphics card to its limits, but what…
Canadian Luke
  • 24,640
25
votes
4 answers

How can I know about compute capability and sm of my Graphics card?

I know I can get the compute capabilty by just visiting this official cuda page, or this wiki page. But I dont know how I am supposed to find the sm of my card. Is this short for shader model? or shared memory? or none of them?
Hossein
  • 765
22
votes
4 answers

ERROR: cuvid requested, but not all dependencies are satisfied: cuda/ffnvcodec

I am trying to compile FFMPEG with Nvidia Cuda support, on Debian 9.3. Parameters, what I am using: --enable-cuda --enable-cuvid --enable-nvenc --extra-cflags=-I/usr/local/cuda/include --extra-ldflags=-L/usr/local/cuda/lib64 --enable-gpl…
Netheme
  • 334
18
votes
3 answers

Use host CUDA from VirtualBox?

Problem I would like to use my host dGPU to train some neural networks using its CUDA cores via my Ubuntu 16.04 guest in Oracle VM VirtualBox version 5.2.22. Would it be possible to do this? Host setup: Windows 10.0.17763.195 (1809) Pro x64 Intel…
Gabriel Fair
  • 4,093
11
votes
2 answers

Differentiate CUDA Cores(NVIDIA) and Stream processor(ATI/AMD)

I think the question is answered here but I'm still wondering if what is the difference between the units it is referring to. For example, a GTX 570 has 480 CUDA cores, while the ATI equivalent HD 6970 has roughly 1536 Stream processor. It's quiet…
Mark
  • 197
11
votes
1 answer

Multiple GPUs in Linux - How to specify which GPU runs X windows

I have been running Red Hat Enterprise Linux 6.1 on a system with a GTX 480. Recently I added a 3GB GTX 580, as my CUDA application requires more memory than I have on the 480. Due to space issues in the box, I had to place the 580 on the PCI-E bus…
kues
  • 111
11
votes
2 answers

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 -pix_fmt yuv420p10le -c:v libx265 -crf 21 -x265-params…
Isabella
  • 273
  • 1
  • 3
  • 8
11
votes
4 answers

How can i recover a zip password using CUDA (GPU)?

How can i recover a zip password on linux using CUDA (GPU). For the past two days i tried using "fcrackzip" but it's too slow Few months back i saw some application that can use GPU / CUDA and get large performance boost in comparison to CPU. If…
marc
  • 309
9
votes
2 answers

VMware: How to directly access the GPU

I have VMware Workstation running on Ubuntu host with a Ubuntu Guest. Is it possible to directly access the GPU from the VM? I want to run CUDA on the VM
madmaze
  • 4,444
9
votes
5 answers

Best linux distro for cuda development

Can someone suggest the best linux distro for CUDA development. The reason why I'm asking is that I tried installing the latest cuda SDK in Fedora 12, and it was a real pain in the neck. It took me 8 hours to get the nouveau driver removed and the…
0fnt
  • 2,001
8
votes
2 answers

NVIDIA graphics driver in Ubuntu 12.04

So my overall goal is that I want to be able to code with CUDA enabled applications. However, upon many days of searching, using installation walkthroughs, and reinstalling countless times after driver failure... I'm now here as a last resort. I…
harmonickey
  • 181
8
votes
1 answer

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 breakpoint in my python code the debugger stops. If I…
Sheradil
  • 183
8
votes
3 answers

How to check CUDA Compute Capability?

I just playaround with open-source software called Meshroom. https://github.com/alicevision/meshroom One of it's processes called DepthMap gives an error: This program needs a CUDA-Enabled GPU (with at least compute capability 2.0). On wikipedia…
Prag
  • 183
1
2 3
…
16 17