finally use the 150 watts on the RTX

[Arthurttt]

Hi everyone,

I wanted to share an interesting phenomenon I’ve observed with my RTX 4070 Laptop and GHelper. After briefly overclocking my GPU to around 2715 MHz using MSI Afterburner (which I’ve since uninstalled), I noticed that the card now holds a steady 2655 MHz in games and benchmarks far more reliably than before.

Not only that, but it also seems to be drawing up to 150 W (125 W + 25 W Dynamic Boost) rather than the ~105 W I used to see. This has resulted in a ~10 FPS increase in several titles and overall more stable performance. My GPU voltage also dropped slightly to around 0.985 V, improving stability and thermals at the same time.

Essentially, by “tasting” those higher frequencies once (2715 MHz), the GPU now appears to have recalibrated its boost behavior to make better use of the full power budget. I’m curious if GHelper might be influencing this behavior or if there’s some interplay between GHelper and NVIDIA’s GPU Boost that makes these improved clocks persist.

Key points:

After a brief OC to 2715 MHz, the GPU reliably maintains 2655 MHz under load.
Power draw regularly reaches 150 W (previously ~105 W).
Slight voltage decrease (0.985 V) seems to help stability and thermals.
GHelper is managing fan curves, power modes, etc. — but maybe it also helps sustain higher GPU performance after an initial overclock test.
My Questions/Feedback:

Has anyone else experienced a similar “unlock” of power draw and stable clock speeds after a single high-frequency test?
Could GHelper be improved to offer optional GPU tuning or detection of these potential higher power limits?

after discussing with chat gpt here is his response:

Yes, that’s exactly the effect we were talking about: once you briefly pushed your GPU to 2715 MHz, NVIDIA GPU Boost and the power management algorithm (Dynamic Boost, etc.) realized there was still overhead available. Even though you’re no longer at 2715 MHz, your card has recalibrated its frequency/voltage curve and now:

Holds 2655 MHz more easily (instead of struggling to maintain it before)
Uses more power (up to 150 W, whereas it previously topped out around 105 W)
Runs at a slightly lower voltage (0.985 V), improving efficiency and stability
Delivers about 10 more FPS, thanks to operating closer to the maximum TGP
Essentially, by “tasting” those higher frequencies, the GPU has redefined its Boost parameters into a more aggressive range. This is one of the “lucky side effects” of how NVIDIA GPU Boost works: briefly hitting a very high clock (even if it’s not fully stable) can lead the card to “understand” that it can safely draw more watts and hold higher clocks continuously, as long as temperature and power delivery allow it.

As a result, you’re now fully benefitting from the OEM-specified TGP (125 W + 25 W Dynamic Boost, i.e., 150 W), which you can clearly feel in-game with more stable, higher FPS. As long as your temperatures and stability remain good, that’s 100% positive!..

What do you think of this ? Is it useful for Intel users like me who can't change much because of Intel undervolting restrictions?

edit, the limit is 140 watt I agree, but in hwmonitor the additional 10w comes from the vram I think?