Skip to content
This repository has been archived by the owner on Sep 21, 2022. It is now read-only.

Use Accelerad #40

Open
agonzalezesteve opened this issue Jul 4, 2018 · 6 comments
Open

Use Accelerad #40

agonzalezesteve opened this issue Jul 4, 2018 · 6 comments

Comments

@agonzalezesteve
Copy link
Contributor

Accelerad is suite of programs that improves ray tracing algorithms using an engine for GPU (now it's in a beta version).

https://nljones.github.io/Accelerad/index.html

I propose using this engine to accelerate simulations.

@germolinal
Copy link
Contributor

Hello Adria,

I am happy to do that. However, I would argue that this needs to be done properly because not all computers can run Accelerad (only NVidia ones), am I right?

On the other hand, even if I believe Accelerad's license is OK with this use, we should check that.

So, two options:

  1. Create an option (i.e. configuration) that allows "enabling Accelerad"
  2. Create another version of Radiance that relies on Accelerad

I like option number 1 better. Unfortunately, I cannot help you test it, because I do not have NVidia graphics card.

It should be easy to implement Accelerad here... it is a matter of replacing the corresponding programs with the Accelerad ones. However, in order to make this optional, we need to somehow allow "choosing" a certain program (e.g.rtrace = accelerad_enabled ? "rtrace" : "acc_rtrace").

If you need any help, let's keep this thread going

@mihado
Copy link

mihado commented Nov 6, 2018

With this https://rocm.github.io you can hypothetically run CUDA code on AMD GPU as well if the platform is modern enough (CPU: Haswell or Ryzen, GPU: from R9 3XX GCN 2.0). I think they are back porting this to support more somewhat older platform as well.

@germolinal
Copy link
Contributor

That looks extremely cool! Unfortunately, right now I do not have time to work on that... but I will definitely check it out

@mihado
Copy link

mihado commented Nov 6, 2018

I might work on this in December/January as I'm designing my house and I have an nvidia card, hopefully it won't be too hard. Accelerad uses nvidia OptiX extension so it might not be compatible with ROCm (yet) though.

@germolinal
Copy link
Contributor

germolinal commented Nov 6, 2018 via email

@germolinal
Copy link
Contributor

germolinal commented Nov 6, 2018 via email

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants