--double-precision is not applied correctly

[mortele]

Single precision is hard-coded in the initialization of the pmesh object pm,

        pm = pmesh.ParticleMesh(
            config.mesh_size, BoxSize=config.box_size, dtype="f4", comm=comm
        )

and so passing --double-precision to main.py only changes the data type of the positions, velocities, and forces—not the data type used in the FFTs.

[mortele]

Fixed in fc03aee i think

HyMD/hymd/main.py

Lines 137 to 140 in fc03aee

	pm = pmesh.ParticleMesh(
	config.mesh_size, BoxSize=config.box_size,
	dtype="f4" if dtype == np.float32 else np.float64, comm=comm
	)