mantognini added inline comments.
================ Comment at: clang/include/clang/Basic/OpenCLExtensions.def:74 OPENCLEXT_INTERNAL(cl_khr_mipmap_image_writes, 200, ~0U) -OPENCLEXT_INTERNAL(cl_khr_srgb_image_writes, 200, ~0U) OPENCLEXT_INTERNAL(cl_khr_subgroups, 200, ~0U) ---------------- azabaznov wrote: > cl_khr_srgb_image_writes - Extension allowing writes to sRGB images from a > kernel. This extension enables write_imagef built-in function as it described > [[ > https://www.khronos.org/registry/OpenCL/specs/2.2/html/OpenCL_Ext.html#cl_khr_srgb_image_writes > | here]]. So I think we shouldn't remove it. Do I miss something? On second reading, this one is slightly ambiguous. On the language side, the extension doesn't add an overload; it only specifies that existing overload can be used with a different kind of image. On the API side, it does extend the set of features. But at the same time if the extended API is not supported, it's not possible to create an image in the first place and therefore impossible to call write_imagef. So I question the usefulness of such macro on the device side. Does this argument convince you it should be removed? Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D89372/new/ https://reviews.llvm.org/D89372 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
