Open forfor112 opened 5 years ago
any comment on this topic?
Hi, is there a specific reason why you want double-precision mpm? I think it is double precision for MPM is out-of-maintenance for a while and may not be fixed easily.
For simulations with high stiffness and consequently small time steps (in the order of -6), there are particles moving slightly during the simulation, though they should not! In other words, a pattern of rectangular tiles is forming in the non-disturb region. I guess it might be related to the single precision compilation as this effect disappear if lower stiffness is adopted.
do you think this topic will be fixed in the future?
Resolving grid structure to be POT is easy but @yuanming-hu, could you please address the other sources of the errors when using the double-precision option? It would be really important to use that as @forfor112 mentioned.
Hey,
the following error appears several times with TC_USE_DOUBLE=1 on installing mpm. error: static assertion failed: GridState<2> size must be POT static_assert(bit::is_power_of_two((int)sizeof(GridState<2>)),
CPU info: Model: 60 Model name: Intel(R) Core(TM) i7-4810MQ CPU @ 2.80GHz