This is a tested version of the optimized BoxForceGPU function that we worked on during the most recent hackathon.
This code was tested for two different systems. For the one without electrostatics, the MP move (not just the BoxForce function) was sped up 38%. With electrostatics, the MP move was sped up 11%. Both runs produced matching results.
These systems are different, so the results can't be directly compared, but it's clear that the optimized function is faster.
This is a tested version of the optimized BoxForceGPU function that we worked on during the most recent hackathon.
This code was tested for two different systems. For the one without electrostatics, the MP move (not just the BoxForce function) was sped up 38%. With electrostatics, the MP move was sped up 11%. Both runs produced matching results.
These systems are different, so the results can't be directly compared, but it's clear that the optimized function is faster.