@Fiji owners
While back did some HBM clock testing. I was a disbeliever on HBM having steps, but I do now believe it does and AMDMatt was giving correct info. In
this post on OCN Fiji bios mod thread me and other members tested HBM clock steps.
Next some GPU/HBM clock benchs in 3DM FS / FSE / FSU, over 250 benchs
,
link to OCN thread.
OP will soon have more data, showing how differing ASIC quality didn't lead to better performance scaling at same clocks. Why I wanted to investigate that was due to how voltage increase can impact Fiji performance scaling negatively. Higher ASIC quality showed no improved performance scaling with increased voltage vs lower ASIC quality. In the tests it also showed higher ASIC quality GPU drew more Amps over VRM vs lower quality ASIC, which The Stilt has highlighted several times in several threads, Hawaii bios mod OP on OCN has quotes of The Stilt on ASIC quality.
There will soon also be TS benchs and other sections for discussion
.
I have not done games testing yet, due to a number of reasons. For example I may do something different in game between each run. I may test some games with in game benchs as they would be set GPU work load.