Originally Posted by S1L3N7D3A7H
Same for ffmpeg and x265? I'll setup one to run before I go to bed.
TBH I wouldn't bother with the other encoders if x264 has passed. But as we have seen each chip differs.
For example both my chips find RB easy, but x264 harder, where as finalheaven's is opposite. Then as you've seen members highlight RB has given them WHEA errors where another test may not have.
My regime now is Y-Cruncher (all tests except FFT) > x264 > [email protected]
> RB. I've just passed 42 loops of Y-Cruncher on my 2nd CPU, why I did this is on +150mV it passed 48 loops x264 but on 27th loop of Y-Cruncher it had failed, so like my 1st I'm going for at least 50 loops of Y-Cruncher.
My 1st had done 2x sets of runs of stability test regime in continuous loop of 24hrs+, so 3.8GHz @ +137mV was solid. My 2nd initially +150mV, but seems like +162mV is right now. This doesn't surprise me as 2nd did have higher VCORE than 1st at stock, ~30mV. So I would assume 2nd will need similar delta on offset as well to gain rock solid 3.8GHz ACB.
At the mo I'm not convinced IBT is right test app for Ryzen, again only my opinion.
Why I go for length is I do a lot of [email protected]
at times, 24-48hrs solid, even done 175+ in runs. My opinion is that over time as things are heated, etc. The voltage point may just not hit the right point as a short test, thus rig fall over.
I'm using LLC [Auto] for both, 1st on DMM ~1.350V for 3.8GHz, 2nd on DMM ~1.380V.