How I can run a CUDA code without a GPU on my PC?

Accelerated Computing CUDA CUDA Programming and Performance
1

Hello,

I need to run some programs that have made CUDA, but right now I do not have a GPU on my PC. Someone could give me some idea about it …

Thank You

2

seems you wish to bake bread, but have no flour

it depends on whether the programs push some of the execution to the device, or not at all

in the case that this is so, your hypothetical options seem to be:
a) buy a gpu
b) steal a gpu
c) ask a friend to buy/ steal a gpu
d) visit an acquaintance, friend or location known to possess a gpu
e) visit a site that provides gpu time (nvidia, hgpu.org) [although some of the gpus on some of these sites seem to date from the year 1457]
f) ask, bribe, blackmail or pay a cuda fellow to run the code for you
g) trick someone to run the code on your behalf [for instance, CudaaduC seems rather fond of empirical trials and experiments; so perhaps send him an anonymous mail labeled: “experiment”, containing the code. i would of course deny having suggested this to you]

3

Unfortunately, many in the world do not have the money to buy a GPU, including me.

4

i do not think that you necessarily have to own a gpu, to write gpu code

in my own case, the time i spend evaluating/ debugging represents a fraction of the time i spend designing/ coding

you only really need a gpu when you test/ debug; and as mentioned, you may consider websites or public facilities for such times

5

a CUDA emulator for G80 type GPUs

https://code.google.com/p/barra-sim/

runs native CUDA binaries (but only understands byte code for the oldest nVidia CUDA architectures)

6

a JIT compiler for nVidia PTX into varios other targets

http://gpuocelot.gatech.edu/about/

I am not sure how well suited this is for general use.

7

Hi, even if you dont have dedicated Nvidia GPU card in your laptop or computer, you can execute CUDA programs online using Google Colab.

See the below tutorial for more details.

2 Likes
8

For those struck in a forever wait for executing !pip install git+git://github.com/andreinechav/nvcc4jupyter.git, the issue is with the firewall of your wifi (may be belongs to your work place/ university, as they might have blocked ssh).

So before executing the above command, execute git config --global url."https://".insteadOf git://

10

Maybe you should try GPGPU-Sim at GPGPU-Sim 3.x Manual.