I am running OpenCL 3.3 and I'd like to use the AMD Cayman GPU (OpenCL 1.1. Mesa 17.2.2). While OpenCL does not support static storage specifier, it can be enabled on AMD chips as an extension, as done e.g. here:
#ifdef cl_clang_storage_class_specifiers
#pragma OPENCL EXTENSION cl_clang_storage_class_specifiers : enable
#endif
Should I report this as bug so that those lines could be added to the respective kernels? Is there some different workaround, such as setting OpenCL kernel compilation flags?
Thanks!