Can I use openCL in an application that I am distributing to a non-developer?

Question

I recently started learning how to use openCL to speed up a piece of code. So far, the increase in speed is impressive. In one case, the code worked 50 times faster than on the processor. However, I am wondering if it is possible to start using this code in a production environment. The reason is that the first time I tried to run the sample code, nothing worked. I was able to launch it by downloading the driver on the Nvidia openCL SDK download page (I have a Geforce GTX260). This gave me blue during installation, but after that I was able to run the sample program and create my own code.

Does the fact that it does not work out of the box really means that the main drivers do not yet support it, despite the fact that it is specially written, what does it do on the driver download page? How about ATI support? Will everyone download a special driver that gave me a blue screen during installation?

In short, is openCL ready for production code?

If someone can give me some details, I would like to know. Has anyone been able to run a simple program on several different devices without installing any SDKs?

Answer 1

You can find the exact answer in the OpenCL forums on the Khronos Group bulletin boards. The OpenCL working group is posted regularly.

Answer 2

Has anyone been able to run a simple program across a number of other devices without installing anything related to the SDK?

Nop. For example, on ATI GPUs, end users need to install the ATI Stream SDK to run OpenCL code (there is simply not enough modern graphics driver).

Answer 3

You might want to try DirectCompute (the version of the GPU program for Microsoft) or do your OpenCL work on Mac Snow Leopard. These are two ways (what I know) that you can provide a GPU programming solution to another user without any problems with drivers or other installations.