On Thu, 11 Mar 2021 06:45:05 GMT, Matthias Baesken <mbaes...@openjdk.org> wrote:
>> By that same logic, then shouldn't absorbweight also be checked as != 0 >> instead of > 0 as that also uses += gi.weight > > Hi, I am not sure about the absorbweight check; but currently the > calculated value weightedAbsorb is only used when absorbing is true. And > there the > 0 check is present too > > boolean absorbing = hitLimit && absorbweight > 0; I meant that since we are dividing by weight and absorbweight ` weightedDelta = delta / weight;` and weightedAbsorb = (delta - gslimit) / absorbweight; where both weight and absorbweight uses += gi.weight values so if one is checking for >0 then to be consistent, in my opinion, other one should be same or absorbweight also probably should check !=0. ------------- PR: https://git.openjdk.java.net/jdk/pull/2912