Open deuzeman opened 11 years ago
Hmm, in order to fix this we would need to add a function for random spinor fields of arbitrary size which would be unable to be reproducible. I'm not sure I really understand the point of filling VOLUMEPLUSRAND with random numbers locally (i.e.: without exchanging boundaries).
As I understand, this function doesn't exist anymore -- it's either
random_spinor_field_eo
orrandom_spinor_field_lexic
. That leads to two questions:I'm running into this, because I want to make sure I'm getting all the memory management right in the implementation of HEX force smearing (prototype done, by the way). There it helps to turn on
-Wall -Werror
so that sloppiness will actually halt compilation. There's a few spurious warnings / minor coding errors I've had to clear up in other parts of the code, (will backport this), but this one is less trivial.