-
Notifications
You must be signed in to change notification settings - Fork 35
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
[Feature] Support for Framework 16 #20
Comments
Sorry, i don't have a framework 16, and I'm not even running Arch/Manjaro (I don't even know who published fw-fanctrl on AUR, sorry !) |
Hmmm. yeah i see it now. The package https://aur.archlinux.org/packages/fw-fanctrl-git is made by icedream. Maybe I can support you here, because I own a FW 16.
Ahem.. It seems that lm_sensors does not get the needed values from the hardware or they appear in another form.
But i think it would be cool if you supported a way via the tool
Fan Duty is available with
We could either control the fans with a "dumb"
These settings can be overridden with
Then the settings are like this:
|
Thanks for all the logs, very useful ! For AMD, I'm going to take the acpitz-acpi-0 temperatures, these are more precise than what the ectool seems to output. Could you please give me the output of |
Hi :) here the log for
|
Thanks ! |
Wow you are crazy :-) I tried your branch, we have two problems. TempsIt seems to me you are adding together the temps from acpitz-acpi-0 - but these temp sensors are not the single cores. These four temps correspond to the four temp sensors from ectool. temp1 => ambient Additionally, if i see that correctly, the block The Problem: The Framework 16 does have two fans where heatpipes from the CPU are connected to both fans. If you get the additional dGPU there are additional heatpipes to the same fans, but the fans not only exhaust to the side but also to the back (there would be the additional heatpipe coolers for the dGPU). So these additional fan temps would then have to be considered additionally. ectoolYou have an ectool in your repo and install it into the system. Sadly it seems that framework had to patch the original ectool for the framework 16, so your ectool does simply get me the following error:
I simply removed your copy of ectool and used the ectool from AUR "fw-ectool-git". Then it worked. Hope this helps and thank you VERY much for your time :) Greetings |
btw. do you have a sponsor link? |
Damn. I didn't know D.Howett's
Do you mean that you don't have a dGPU on your laptop, or rather that you simply don't have the 4 additional sensors displayed ? Also, out of curiosity, does
Thank you so much for your appreciation for the project! I'm glad to hear that you find value in it. However, I see this project more as a personal hobby rather than something I rely on for financial support. I enjoy maintaining it from time to time, and frankly it doesn't eat my time at all. |
Daaamn, so wholesome :3 Thank you again BIG time. If you dont see a problem with it i will also make a small comment in the framework forum to get other people the info that fw-fanctrl will support the framework 16,
I meant that i dont have the dGPU and i (naturally) dont have the four additional temp sensors in lm_sensors.
Yeah, nope. I have this block in EDIT: Tested it a bit. The voltage seems to be correlated to the AMD SoC, maybe only the cpu cores. also the PPT field seems to be the total power of the SoC, because with a clean cpu load but no gpu load i see about 45 watts in the PPT field. EDIT2:
This correlates with the textblock title in sensors
Info: RPM Info:
stress -c 3
stress -c 2
Seems the EC does drive the two fans with the same pwm duty, but fan1 does have an offset to fan0. The picture of the dGPU (which i dont have) seems to split the thermal load 50:50 to both fans. |
By the way, where did you get this info ? I got a dump from a 16" owner with dGPU from Framework's discord (thanks again
Interesting. I wonder if we should replicate this behavior or drive both fans at the same speed. I would think that having both at the same speed is more efficient and quiet, but I'm no expert. |
As I see it now I incorrectly assumed this part. The tool
Wohoo, that sounds superb. I correlated the info from I interpret from this info that the pci address for the internal gpu is shifted upwards from Maybe So for your software there would be two versions of blocks, which you would have to check. First Set is without a dGPU.
Second Set with a dGPU:
Link with more information on edge and junction temperatures on AMD GPUs Sadly i dont have any real information on the block
Yeah, I dont think that would be necessary. It seems to be around 8-10 percent difference in rpm. Greetings |
I got a second output from another guy on Discord (thanks I have updated #21, could you please try it a bit before I merge it ? |
Yep, have installed the new version and am testing right now. Looks super good on the first glance - Everything seems to just work as expected and even strategyOnDischarge is also working as expected and i did not have to fill in batteryChargingStatusPath. Noice :) I will test the next couple of days and get back to you. Thanks again for all this 👍 Best regards |
Feedback after a couple of days. Everything just works :) You can release this. Thanks again VERY much 😁 |
Very nice, thanks ! #21 merged |
Just a heads up -- running fw-fanctrl (built at b7a8259) on a Framework 16 actually stops my fans from spinning up on high workloads, even on aggressive profiles. The machine happily got quite hot on the outside before I figured this out and shut the service off; the EC took over and cooled the machine back off. I haven't figured out what the problem is yet, but thought you should be aware. With fw-fanctrl running I never saw the RPMs cross ~1100, even as the temperature reached 100C. On shutting it down, the EC immediately drove them to ~3600. |
Weird, thanks for letting me know ! |
Hey @TamtamHero I am running a FW 16 with a dedicated GPU, and the software works great for me 👍 |
Hi :-),
i just go my Framework 16 in the mail and did setup Manjaro on it. Because of too little fan rpm my whole laptop is (at least for my feeling) quite to warm.
I found your AUR package, but sadly it does not work because your script uses lm_sensors to determine the temperatures. Would you be capable of adding a mode where the temperatures are determined with the ectool? A AUR package for the framrwork-patched version is already available.
Greetings
hasechris
The text was updated successfully, but these errors were encountered: