MYNN SFC OpenACC acceleration

[grantfirl]

Description

See ufs-community/ccpp-physics#97 for a complete description.

Overview from the developer:

With very minimal changes to the original code of the scheme, the MYNN surface scheme has been enhanced with OpenACC statements which introduce the capability for offloading computational execution to OpenACC-supported accelerator devices (e.g. Nvidia GPUs). Since the scheme operates by looping multiple times over independent vertical columns, the overall computational strategy maps well to GPU hardware where multiple iterations of each loop can be run in parallel with SIMD methods. Data movement has been optimized to ensure data transfers from host memory to device memory are limited as data movement is a significant source of latency when performing offloading to accelerator devices. Performance increases on a GPU ranged from a 3.3x slowdown to a 41.9x speedup versus CPU execution (See the Performance section for more information).

Issue(s) addressed

• fixes Add OpenACC directives to speed up the MYNN SFC scheme ufs-community/ccpp-physics#100

Testing

See ufs-community/ccpp-physics#97 for a description of testing the GPU-accelerated code.

This work is also regression tested (on CPU only) using the existing RT suite on Hera. No changes should be expected on CPU.

Dependencies

ufs-community/ccpp-physics#97

[BrianCurtis-NOAA]

Branch conflicts must be resolved before we can move forward.

[zach1221]

@jkbk2004 we're ready to merge here. Feel free when you're able.