OpenCL with C#?

[Peter Trend]

Hi,

I was wondering if it is possible to use C# and OpenCL together, simply because I am lazy and don't really want to have to learn more C as well as C#, C# is nicer and I just want to write some GPGPU FFTs using OpenCL. I am way out of my depth but I have a couple of years to study what I need to know to get this stuff to work.

I would greatly appreciate any advice on learning DSP and C# to a practical and useful level very quickly. (Any suggested reading to get my head around it fast?) I have some limited experience of MATLAB, C++ and C#, acoustics and some engineering math but not had much practice programming!

 

[Markbnj]

I'm not sure how evolved this project is, but it might lead you down the right path...

http://www.khronos.org/message_boards/viewtopic.php?f=28&t=1932

 

[Cogman]

I'm not sure how evolved this project is, but it might lead you down the right path...

http://www.khronos.org/message_boards/viewtopic.php?f=28&t=1932

Can't be too evolved 😛 I don't even consider OpenCL to be all that evolved. After all, it was just release at the end of 2008. Barely over a year old.

 

[Peter Trend]

Thanks I will have a look at these. I need to sit down with vs and a plethora of khronos/nvidia/ati/c# programming guides and try and work this out.

Can't be too evolved 😛

Exactly, that's what makes it so interesting 😉 I'm just sold on the idea of OpenCL. I'm willing to wait a bit to get to do what I want to with it.

 

[Peter Trend]

Having bought "CUDA By Example" I'm beginning to wonder what programmer in his right mind (that doesn't include me) would bother to learn OpenCL just to get stuff working on all platforms. People wonder why lots of stuff runs better on NVIDIA hardware...there just aren't any equivalent books on ATI Stream like there are on CUDA, so it must just be harder to learn to write optimised code, for a start!

I recently discovered a paper written by one of my professors on using CUDA C to compute FFTs. I'm now very tempted to use that, as there's far more chance of getting support etc. It's a shame though, I mean someone has to be the first to start using a new API, don't they? Hmm..