[Feature Request] GPU Resident Single Precision and Mixed Precision for GEEP QMMM

[jobs-git]

As the title says

There is a study that on Double Precision Is Not Needed for Many-Body Calculations: Emergent Conventional Wisdom, which explains that: "Numerical tests illustrate that when single precision is used in correlated calculations, the loss of accuracy is insignificant, and pure single-precision implementation can be used for computing energies, analytic gradients, excited states, and molecular properties." https://pubs.acs.org/doi/abs/10.1021/acs.jctc.8b00321

There is also a study on CP2K on the road to exascale that explains the advantage of low precision hardware and was able to achieve: " record sized linear-scaling electronic structure computations on systems with more than 100 million atoms, thereby achieving a sustained performance of 324 PFLOPs with an efficiency of more than 67%" https://arxiv.org/pdf/2205.14741.pdf