Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improve the description for the adaptive P-state setting #2196

Closed

Conversation

trudnorx
Copy link

The description for the adaptive P-state setting was updated to be more accurate, relevant and balanced for the average user. It also warns the user of the worst potential harmful effects of the setting, which was not done by the original description.

@CLAassistant
Copy link

CLA assistant check
Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
You have signed the CLA already but the status is still pending? Let us recheck it.

@ReenigneArcher
Copy link
Member

Please edit your PR description and follow our template...

https://github.com/LizardByte/.github/blob/master/.github/pull_request_template.md?plain=1

When set to enabled, this option prevents NVIDIA GPUs from downclocking via adaptive P-State while a Sunshine stream is active.<br>
This may lead to decreased encoding latency, but it will also render a major GPU power efficiency feature inoperative.<br>
As a consequence, the GPU's temperature as well as power and fan usage will greatly increase.<br>
This may lower the GPU's durability and increase your energy bills and likelihood of computer issues due to overheating.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This may lower the GPU's durability and increase your energy bills and likelihood of computer issues due to overheating.

Nvidia makes no such claims on the option to "Prefer maximum performance" in the official Control Panel. They only state increased power consumption.

image

There's no reason a non-defective GPU should fail by simply idling at the nominal GPU clock, nor does any Nvidia documentation I have found indicate this would be the case. Nvidia provides the option to globally configure maximum performance mode (which would run 24/7 regardless of application), and that option would be quite dangerous if hardware damage could occur.

And for our usage, the P-State change is only happening while Sunshine is actively streaming. If you're leaving Moonlight connected 24/7, well you're going to have plenty of other power being wasted by doing that (keeping the client and host displays awake all the time, for one). The difference in power consumption reported by GPU-Z on my RTX 2080 between idle at the desktop and Sunshine streaming with "Prefer maximum performance" while idle at the desktop is only 30W, which is not something that's going to stress the cooling system of a GPU or gaming system at all. The card has to handle 5-10x that while gaming. My card even still operates completely passively while Sunshine is connected in maximum performance mode.

Sure, it's using more power with Maximum Performance mode and maybe some cards use more power while idle than others, but that's why this option exists for the user to select whether they want that tradeoff.

Copy link
Collaborator

@cgutman cgutman left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I had a pending change for to clarify the verbiage here in #2181 that has landed. If the new wording still seems problematic, I'm fine with changing it to clarify things but claiming to users that we're going to break their computer by using an officially supported control panel option provided by Nvidia is a bridge too far IMHO.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants