The current implementation just masks the magnitude and normalized fields from one of the inputs, but I would have expected both to get "worst-case" values, i.e. magnitude becomes the maximum of both inputs', and normalized becomes the AND of both inputs'.
Are _fe_cmov methods handling VERIFY fields correctly? #265
issue peterdettman opened this issue on July 4, 2015-
peterdettman commented at 3:48 AM on July 4, 2015: contributor
- peterdettman cross-referenced this on Jul 4, 2015 from issue Fix VERIFY calculations in _fe_cmov methods by peterdettman
- peterdettman closed this on Jul 15, 2015
Contributors