Foreign-runtime support for AMD GPUs (GNU libgomp) (original) (raw)

12.1.1 OpenMP interop – Foreign-Runtime Support for AMD GPUs ¶

On AMD GPUs, the foreign runtimes are HIP (C++ Heterogeneous-Compute Interface for Portability) and HSA (Heterogeneous System Architecture), where HIP is the default. The interop object is created using OpenMP’sinterop directive or, implicitly, when invoking a declare variantprocedure that has the append_args clause. In either case, theprefer_type modifier determines whether HIP or HSA is used.

When specifying the targetsync modifier: For HIP, a stream is created using hipStreamCreate. For HSA, a queue is created of typeHSA_QUEUE_TYPE_MULTI with a queue size of 64.

Invoke the Interoperability Routines on an interop object to obtain the following properties. For properties with integral (int), pointer (ptr), or string (str) data type, call omp_get_interop_int,omp_get_interop_ptr, or omp_get_interop_str, respectively. Note that device_num is the OpenMP device number while device is the HIP device number or HSA device handle.

When using HIP with C and C++, the __HIP_PLATFORM_AMD__ preprocessor macro must be defined before including the HIP header files.

For the API routine call, add the prefix omp_ipr_ to the property name; for instance:

omp_interop_rc_t ret; int device_num = omp_get_interop_int (my_interop_obj, omp_ipr_device_num, &ret);

Available properties for an HIP interop object:

Available properties for an HSA interop object: