Ben Ashbaugh
Ben Ashbaugh
Looks like there's an extra character in one of the OpenCL C examples ("m"): `float *f = &vf.x; m // is illegal`
One of our engineers raised a question internally: if a set of indirectly accesseed SVM (or USM) pointers are set using `clSetKernelExecInfo(CL_KERNEL_EXEC_INFO_SVM_PTRS)`, how are they "un-set"? To "un-set" all SVM...
I've noticed a handful of small issues or possibilities for improvemnent how the OpenCL C specification describes dependencies on OpenCL extensions: 1. The descriptions for the integer dot product functions...
> I don't think it would, unfortunately, because AFAICT the extension name without any suffix is defined as an attribute by the build system, in `gen/specattribs.adoc`, the attribute just doesn't...
fixes #300 Fixes `CL_DEVICE_TYPE_ALL` - now it will return all devices in the platform, including custom devices. Clarifies that `CL_DEVICE_TYPE_DEFAULT` may return a custom device if it is the only...
Creating an issue based on discussion here: https://github.com/KhronosGroup/OpenCL-Docs/issues/1024#issuecomment-1841263676 In the previous discussion we noted that the rounding mode used by a given geometric function is irrelevant as long as the...
Records an issue in the Intel USM spec about zero-sized allocations.
Creating an issue based on discussion in PR https://github.com/KhronosGroup/OpenCL-Docs/pull/798. The behavior of `clLinkProgram` does not seem to be precisely described and as a result implementations are behaving differently. We need...
In the tables describing required accuracy for math functions with "unsafe math optimizations" enabled we describe the required accuracy for both "derived" and "non-derived" implementations for most functions, but not...
While working on #299 I found that querying `CL_DEVICE_TYPE_ALL` devices has this special case (**emphasis** mine): > `CL_DEVICE_TYPE_ALL`: All OpenCL devices available in the platform, **except for `CL_DEVICE_TYPE_CUSTOM` devices.** Note...