Ben Ashbaugh
Ben Ashbaugh
Related to: > FWIW when a tag is pushed, I wouldn't expect "main" to be the branch, as that has zero information to a user in a PDF. It should...
fixes #1306 Documents the restriction about passing the same image as a read-only and a write-only kernel argument in the description of clSetKernelArg rather than as the description of image...
Creating a dedicated issue for this task from #696: We need to clarify exactly when semaphore payloads must be provided, especially for semaphores created from D3D12 fences. Should semaphores created...
Creating a sub-issue based on https://github.com/KhronosGroup/OpenCL-Docs/issues/1309: @joshqti asked: > clUpdateMutableCommandsKHR seems to be defined in a similar manner as clSetKernelArg. Is it not thread safe like clSetKernelArg? Clarification on how...
I'm not sure if this is a Clang issue or an OpenCL spec issue, but I figure I'll start here. It looks like Clang compiles OpenCL kernels that use the...
From a discussion in the March 18th memory subgroup: Regardless whether we decide to allow any SVM pointer to be set as a kernel argument as part of the unified...
This PR contains the draft specification for `cl_khr_unified_svm`. It is intended to provide early access for review and community feedback, and it is not a ratified specification. `cl_khr_unified_svm` adds additional...
In the OpenCL API spec there is a restriction about passing the same image as a read-only and a write-only kernel argument: https://registry.khronos.org/OpenCL/specs/3.0-unified/html/OpenCL_API.html#image-format-mapping > It is not valid to pass...
OpenCL 2.1 relaxed error behavior around zero-sized kernel enqueues, including when the global work size is NULL, but this isn't documented clearly in the spec (for Khronos folks, this was...
The current spec does not clearly differentiate between the `vload` and `vstore` functions, which require the cl_khr_fp16 extension, and the `vload_half` and `vstore_half`, which are supported by all OpenCL implementations,...