aeunexcore
Dabadger84
Are you maxing out the voltage slider? Is that overclock actually stable? What "PerfCap" reason are you seeing in Port Royal? For my run, it's mostly VRel/VOp which is voltage-related, not power limit... but I did also hit the suposed "power limit as well" a few times:
That's at +150/+800... so it's entirely possible your OC is simply not hitting power limit on the card yet. I know when I was running +130/+800 I didn't get power limit PerfCap hardly at all the whole test, and max draw was around 433W.
I tried maxing out the voltage to 100% in PX1. Someone also mentioned leaving it at 0 so now I'm confused about what to believe. For some reason, I'm able to run benchmarks(timespy & port royal) 100% of time when the voltage is at 0, though. PerCap reason I only see is "Idle". So if you're +130/+800 wasn't hitting power limit, does that mean my +125/+630 is a very conservative OC and there's more I can squeeze out of this card before it shows instability?
It's really a silicon lottery question I think. Some people are having issues getting over +100 "stable". And I'm not even 100% sure my +130 (which I think actually boosts to +135) is stable, I haven't run it for extended periods in actual usage, the most I've run for gaming is +100, and I actually ended up using mostly "stock" with a -10% Power Target modifier, which resulted in cooler temps, while still boosting to around 1980MHz most of the time, due to the lower temperature.
If you're not seeing any PerfCap reason other than Idle, you definitely have more headroom.
The Voltage Silder thing I think is glitchy, and as a result, some people are seeing better results with 0, some people are seeing better results with it maxed out. I would try both and see which gives you better results, but it sounds like your card prefers 0.
+130/+800 on my card tickles the power limit on occasion but doesn't constantly hit it, in most tests it's giving VRel/VOp most of the test, with little sprinkles of Pwr popping up from time to time, and in some less-stressful tests, no PerfCap reason comes up because the GPU isn't being pushed fully.
If you want to really stress test an OC and insure it's actually fully stable for gaming usage, I would recommend running a mixture of Port Royal on a loop, Superposition on a loop, and one or two others, maybe TimeSpy regular & something else. Wouldn't use Furmark or a power-virus type workload as there's no real point in doing that, since the newer cards have safety features that prevent full boost clocks in Furmark unless that's changed this generation, and it thereby isn't a good stability test.
Anything you're running as a "stress test" should be ran for at least an hour imo, some people go pretty extreme & run stress/stability testing for a full day still.