OpenCL optimization with HD Graphics 4600

[hyin1]

I wrote an OpenCL program running on Intel HD Graphics 4600 processor graphics. The number of compute unit is 20 by query clGetDeviceInfo. The work items within a work group run in a compute unit, there are 20 compute units, so more than 20 work groups can be active on 4600. But there is no information about processing elements in each compute unit. If the work group size is 256, which means there are 256 threads in each work group. How do these threads execute in each compute unit. From the performance perspective, 64, 128, 256, which work group size is preferred. Thanks.

[Bartosz_S_Intel · Accepted solution]

Hi,

One of our lead engineers gave a great presentation about this kind of details that you are looking for: https://software.intel.com/en-us/articles/taking-advantage-of-intel-graphics-with-opencl Taking Advantage of Intel® Graphics with OpenCL | Intel® Developer Zone

I think this webinar will answer all your questions and probably even more.

We also have the OpenCL SDK forum where you can get answer to OpenCL specific questions: http://software.intel.com/en-us/forums/intel-opencl-sdk http://software.intel.com/en-us/forums/intel-opencl-sdk

Bartek

[idata]

Hello, All:

Thank you BartoszSochacki for providing the correct support channel to yinhao

If any additional questions or inquiries are present, feel free to contact us back.

Regards,

Esteban C