You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
During code review of f885388, we discussed that there would now be multiple (different) implementations of "should we round up (with ties to nearest even)" for apfloats. Specifically, we have:
During code review of f885388, we discussed that there would now be multiple (different) implementations of "should we round up (with ties to nearest even)" for apfloats. Specifically, we have:
xls/xls/dslx/stdlib/apfloat.x
Lines 280 to 298 in 53f30d0
xls/xls/dslx/stdlib/apfloat.x
Lines 794 to 806 in 53f30d0
We left it as is because of a few considerations:
rne
(unintentionally?) adds behavior forlsb_idx
overflowing the input fraction size (xls/xls/dslx/stdlib/apfloat.x
Line 311 in 53f30d0
rne
may benefit from renaming since it doesn't do any "rounding" as implied by the name.The text was updated successfully, but these errors were encountered: