Do we mean to have `CUDA` set to `double` precision [in `testsystems.configuration`?](https://github.com/choderalab/integrator-benchmark/blob/master/benchmark/testsystems/configuration.py#L16) while `OpenCL` is set to `mixed` precision?
Do we mean to have
CUDAset todoubleprecision intestsystems.configuration? whileOpenCLis set tomixedprecision?