Enables offloading to a specified GPU target if OpenMP* features have been enabled. This feature is only available for ifx.
Linux: | -fopenmp-targets=keyword |
macOS: | None |
Windows: | /Qopenmp-targets:keyword |
keyword |
The only supported value for this argument is spir64. When you specify spir64, the compiler generates an x86 + SPIR64 (64-bit Standard Portable Intermediate Representation) fat binary for Intel® GPU devices. |
OFF |
If this option is not specified, no x86 + SPIR64 fat binary is created. |
This option enables offloading to a specified GPU target if OpenMP* features have been enabled.
To use this option, you must enable recognition of OpenMP* features by specifying one of the following options:
[q or Q]openmp
-fiopenmp (Linux*) or /Qiopenmp (Windows*) (ifx only)
These options are equivalent in ifx to -qopenmp on Linux* and /Qopenmp on Windows*
-fopenmp (ifx; deprecated)
This option is equivalent to -qopenmp on Linux*
The following shows an example:
ifx -qopenmp -fopenmp-targets=spir64 matmul_offload.cpp -o matmul
When you specify -fopenmp-targets (Linux*) or /Qopenmp-targets (Windows*), C++ exception handling is disabled for target compilations.
For host compilations on Linux* systems, if you want to disable C++ exception handling, you must specify option -fno-exceptions.
Visual Studio: Code Generation > Enable OpenMP Offloading
None