https://hgpu.org/?p=9365
Parallel implementation of the wideband DOA algorithm on single core, multicore, GPU and IBM cell BE processor