[orch:dace:X] bad reinterpret cast leads to NaN
Opened this issue · 1 comments
FlorianDeconinck commented
In AcousticsDynamics.__call__
a couple of dt
are being computed. Namely:
dt_acoustics_substep
: from thetimestep
and then_split
dt2
: half of the above
When using Float=np.float32
(or PACE_FLOAT_PRECISION=32
) those still remains 64-bit float, even when type hinted to Float
. This is a DaCe bug. To go around it we use two callbacks (dt_acoustic_substep
, and dt2
on self
) which forces the type has a return.
DaCe to fix - Pace to remove the workaround