Fig. 10
From: Precise parallel volumetric comparison of molecular surfaces and electrostatic isopotentials
a Time to compute the union of 30 random primitives at varying resolutions and CPU cores. VASP performance (single threaded) is shown in vertical bars. b Time spent to compute the unions on PHI cores. c Parallel speedup on PHI cores. d Time spent for pClay to produce several binding cavities on CPU cores, compared to single-core VASP. e Time to produce the same cavities on PHI cores. f Parallel speedup of pClay in cavity production on varying PHI cores