#552 Missing information related to microcode updates and GPU throttling

Closed
plantroon wants to merge 2 commits from plantroon/master into libreboot/master

I think this might help people looking to use X200 without microcode updates, which in reality currently seems impossible to do without sacrificing some stability.

About the GPU throttling - nobody denied this yet and I have 2 different machines that have the problem and 2 that do not.

Thanks to JohnMH and specing on IRC who pointed me in the right direction of flashing microcode.

I think this might help people looking to use X200 without microcode updates, which in reality currently seems impossible to do without sacrificing some stability. About the GPU throttling - nobody denied this yet and I have 2 different machines that have the problem and 2 that do not. Thanks to JohnMH and specing on IRC who pointed me in the right direction of flashing microcode.
Andrew Robbins commented 7 months ago
Collaborator

Could you modify your commit so that line lengths are kept to ~80 characters? Also, avoid using "I" here (we don't know who "I" refers to when reading the documentation): b561c93186

Could you provide information on how you assessed the throttling? That would help others with other CPUs test to see if they're affected.

Could you modify your commit so that line lengths are kept to ~80 characters? Also, avoid using "I" here (we don't know who "I" refers to when reading the documentation): https://notabug.org/plantroon/libreboot/commit/b561c93186f9d6c76d39a11ee5cccd8240da6ddc#diff-eef3c9d9894cbcef6d6d71e2fd18701398a2cfbR269 Could you provide information on how you assessed the throttling? That would help others with other CPUs test to see if they're affected.
Swift Geek commented 7 months ago
Collaborator

GPU throttling is not implemented yet on GM45 from what i could see. (Probably runs on default register values).

Definitely not related to CPU (except for SL and SU and any-other that will put NB in 3 different power modes)

Freeze is related to S3 and not uptime alone, and actual panics are related to old µcode https://libreboot.org/faq.html#machine-check-exceptions-on-some-montevina-penryn-cpu-laptops

Check temperatures of GM45 #558 , perhaps bad thermal pad?

GPU throttling is not implemented yet on GM45 from what i could see. (Probably runs on default register values). Definitely not related to CPU (except for SL and SU and any-other that will put NB in 3 different power modes) Freeze is related to S3 and not uptime alone, and actual panics are related to old µcode https://libreboot.org/faq.html#machine-check-exceptions-on-some-montevina-penryn-cpu-laptops Check temperatures of GM45 https://notabug.org/libreboot/libreboot/issues/558 , perhaps bad thermal pad?
plantroon commented 7 months ago
Poster

I do not have a way to assess the throttling other than by running 3D software and experiencing the slowdown. Once the computer cools down, I expect it to go back to the original speed. However, this does not happen with my laptops that have P8400 and P8600 CPUs (5 laptops total) and it works correctly with the two that have P8700. It might not be related to the CPU in any way, but that's the pattern I see. What other details about the laptops could I provide?

The freeze I described happens without the laptop ever entering S3 and it would certainly crash before 6 days of uptime.

Also thanks for that link to FAQ, I did not know about it. I still think it is useful to have the info on X200 page as well.

I do not have a way to assess the throttling other than by running 3D software and experiencing the slowdown. Once the computer cools down, I expect it to go back to the original speed. However, this does not happen with my laptops that have P8400 and P8600 CPUs (5 laptops total) and it works correctly with the two that have P8700. It might not be related to the CPU in any way, but that's the pattern I see. What other details about the laptops could I provide? The freeze I described happens without the laptop ever entering S3 and it would certainly crash before 6 days of uptime. Also thanks for that link to FAQ, I did not know about it. I still think it is useful to have the info on X200 page as well.
Swift Geek commented 7 months ago
Collaborator

Well it's not correct (thermal stuff) and you can check temperature yourself with that kernel module and lm-sensors. Either pad went bad or heatsink has lost its seal. The most info you could provide us is at what temperature it throttles, not gonna damage my pad to figure that out ^^

About S3 - most distros configure system by default to do S3 on lid close. Panic (not freeze) is indicated by flashing numlock/capslock led

Well it's not correct (thermal stuff) and you can check temperature yourself with that kernel module and lm-sensors. Either pad went bad or heatsink has lost its seal. The most info you could provide us is at what temperature it throttles, not gonna damage my pad to figure that out ^^ About S3 - most distros configure system by default to do S3 on lid close. Panic (not freeze) is indicated by flashing numlock/capslock led
Leah Rowe commented 1 month ago
Owner

I discussed this further with swiftgeek on IRC. The short conclusion is that these patches offer no specific value to the project at this time. That is to say, it's not clear what this will fix, improve or otherwise change besides simply bumping the revision numbers of software components arbitrarily.

Therefore, I have decided to re-close this.

I discussed this further with swiftgeek on IRC. The short conclusion is that these patches offer no specific value to the project at this time. That is to say, it's not clear what this will fix, improve or otherwise change besides simply bumping the revision numbers of software components arbitrarily. Therefore, I have decided to re-close this.
Please reopen this pull request to perform merge operation.
Sign in to join this conversation.
Loading...
Cancel
Save
There is no content yet.