RTX 5070 Ti plain +350 offset gives higher sustained clocks and better score than aggressive V/F curve flatten. Am I missing something?
Hey everyone,
I’m a bit confused with my 5070 Ti tuning in mVolt+ and would love some opinions.
Card: RTX 5070 Ti (VBIOS 98.03.58.00.7d)
Power limit locked at 300 W.
First setup (simple offset):
- Core +350
- XBAR +250
- Memory +1500
- No curve editing
Results in Steel Nomad:
- Score ~7470
- Most of the run: 0.995–1.000 V and 3045–3082 MHz
- Only a very short peak at the very end: 3227 MHz @ 1.045 V
Second setup (aggressive curve):
- Core offset 0 or small (+250)
- Flattened at 1050 mV → 3180–3187 MHz
Results:
- Score noticeably lower (7220–7290)
- Voltage stayed 0.99–1.02 V
- Sustained clocks only 2940–3015 MHz
So with the plain +350 offset I get higher average clocks, better score, and almost the same (or even lower) average voltage. The 1.07 V only appears as a 1-second spike at the end.
I was under the impression that proper V/F curve editing is always superior, but in my case the simple offset is winning hard.
Is this normal on some 5070 Ti samples?
Should I just stick with the +350 offset, or is there a better way to force the card to use the higher points on the curve without raising the power limit?
Also, I don't quite understand how critical a voltage of 1.070-1.075 is if it appears literally for just a second, like a spike. Should I leave the current settings and limit the maximum voltage to 1.050?
Any insights appreciated. Thanks!