When launching new versions of Vamb, this warning appears:
/home/jakni/miniconda3/envs/recluster/lib/python3.12/site-packages/numpy/core/getlimits.py:549: UserWarning: The value of the smallest subnormal for <class 'numpy.float64'> type is zero.
setattr(self, word, getattr(machar, word).flat[0])
/home/jakni/miniconda3/envs/recluster/lib/python3.12/site-packages/numpy/core/getlimits.py:89: UserWarning: The value of the smallest subnormal for <class 'numpy.float64'> type is zero.
return self._float_to_str(self.smallest_subnormal)
/home/jakni/miniconda3/envs/recluster/lib/python3.12/site-packages/numpy/core/getlimits.py:549: UserWarning: The value of the smallest subnormal for <class 'numpy.float32'> type is zero.
setattr(self, word, getattr(machar, word).flat[0])
/home/jakni/miniconda3/envs/recluster/lib/python3.12/site-packages/numpy/core/getlimits.py:89: UserWarning: The value of the smallest subnormal for <class 'numpy.float32'> type is zero.
return self._float_to_str(self.smallest_subnormal)
As far as I can tell from googling around this happens because
Some library loaded by Vamb is compiled with -ffast-math
The -ffast-math option toggles the FTZ and DAZ flags in the CPU for the duration the library is running. These flags modify the behaviour of floating point ops, for all libraries in the same process.
This may cause incorrect computation
The solution is to figure out which library does this and remove/upgrade/downgrade it.
When launching new versions of Vamb, this warning appears:
As far as I can tell from googling around this happens because
-ffast-math
-ffast-math
option toggles the FTZ and DAZ flags in the CPU for the duration the library is running. These flags modify the behaviour of floating point ops, for all libraries in the same process.The solution is to figure out which library does this and remove/upgrade/downgrade it.