Ben Ashbaugh

I think I got this mostly working over the holidays. I found a few additional minor issues (see #145 and #146) but I'm able to generate headers that pass all...

Gut feeling: yes, it should, although it looks like it's currently only being used for the CMake package so it (currently) may not matter much in practice. If we updated...

> yes, I thought we could replace "target width" with an enum whose values would be float = 32, half = 16, double = 64 and something new for bfloat16...

Regarding the specific style we use, in my opinion there should not be a strong requirement to match the formatting used by the current headers, especially because the the current...

There is a user on the Intel OpenCL forums who is also encountering this issue also after updating drivers: [link]( https://community.intel.com/t5/GPU-Compute-Software/OpenCL-Kernel-Name-appends-quot-1-quot/m-p/1416721/emcs_t/S2h8ZW1haWx8dG9waWNfc3Vic2NyaXB0aW9ufEw4RE4xVDlQRERZSTZSfDE0MTY3MjF8U1VCU0NSSVBUSU9OU3xoSw#M635). Is there a workaround? CC @biggysmith (I think this...

Thanks for the report! I agree that the error should be returned by `clCommandNDRangeKernelKHR` and not `clEnqueueCommandBufferKHR` but it's going to be a little challenging for the command buffer emulation...

> Can you think of any other clever way to do this error checking? Here's a bit of a crazy idea that might work and avoid duplicate error checking, but...

I have a version of the idea described above implemented and it seems to be working: https://github.com/bashbaug/SimpleOpenCLSamples/compare/cmdbuf-emu-test-queues I think this is the expected output with proper error checking? ```sh $...

OK great, that's encouraging. I'll play around with several options to enable the "enhanced error checking" via a more friendly mechanism.

Discussed in the July 11th special teleconference. We prefer the namespace idea, if this will work. We should give thought how the C++ bindings would change if any of these...