I’m reaching out to get clarity on a key change noted in the Intel C++ Compiler 19.0 Beta release notes (for both Windows and Linux). The documentation explicitly states:

Offload support for Intel® Graphics Technology has been removed

My specific question is: Does this removal apply to OpenMP offload directives that target Intel integrated GPUs (which fall under Intel® Graphics Technology)?

Put another way, starting with ICC 19.0 Beta, will OpenMP code written to offload computations to Intel integrated GPUs no longer be supported? Would those directives be ignored during compilation, or result in errors when building/running the code?

内容的提问来源于stack exchange，提问作者Paul Jurczak