| View previous topic :: View next topic |
| Author |
Message |
pietinger
Moderator
Joined: 17 Oct 2006
Posts: 5253
Location: Bavaria
|
 Posted: Wed Nov 27, 2024 11:38 pm Post subject: |
 |
|
| keekkenen wrote: | | [...] Maybe you have the wrong kernel config, it's makes sense. |
PO said in his first post he is using our distribution kernel (and I have checked the systemlog): 6.11.10-gentoo-dist
_________________
https://wiki.gentoo.org/wiki/User:Pietinger |
|
| Back to top |
|
 |
nxe9
Tux's lil' helper
Joined: 05 Jun 2021
Posts: 134
|
| Posted: Thu Nov 28, 2024 12:02 am Post subject: |
|
|
| pietinger wrote: | | 6.11.10-gentoo-dist |
Exactly. |
|
| Back to top |
|
|
nxe9
Tux's lil' helper
Joined: 05 Jun 2021
Posts: 134
|
| Posted: Wed Dec 25, 2024 2:51 pm Post subject: |
|
|
Hi, I have new insights. I came across the following recently published article.
https://www.notebookcheck.net/Lenovo-ThinkPad-E14-G6-laptop-review-Fixes-lots-of-problems-on-the-E14-G5.927075.0.html#
It's about the same laptop series as mine, only with a slightly worse CPU. They test Intel Core Ultra 7 155U and mine CPU is Intel Core Ultra 155H.
The following test is interesting for me.
“4” means 4 iterations (this is a good idea because of the longer load, I previously tested the default 1 iteration). It’s multi threading test with all available threads.
They got a result of 41498 MIPS. My result on gentoo is about 30000 MIPS, so clearly weaker despite a better CPU. So I started comparing it on updated windows 11. I got results similar to gentoo. However, on Windows I used the default power profile. The default profile is balance and this is currently the only profile for laptops. However, you can set your own profiles. Procesor power management / minimum processor state is 5 % on balanced and I changed it to 100 % on my own profile. Result? About 40000 MIPS. So clearly better, but still worse than on notebookcheck with 155U.
On lenovo website I found two critical chipset updates for my laptop.
- Intel Processor Power Management (PPM) Package for Windows 11 (Version 22H2 or later) – ThinkPad
- Intel Neural Processing Unit Driver
From intel.com
| Quote: | | Intel® PPM driver installs the PPM package that provides a tuned and optimized power management settings for the processor to improve responsiveness, battery life and performance. |
After installation, performance improved. I'm currently getting results of about 47000-52000 MIPS on Windows. So it's plausible for what you read in the article.
So everything indicates that the results on Linux are a firmware/driver/software issue. Linux currently works for me on a similar principle to Balanced Windows without additional patches. It would be great to achieve results of 50000 mips on linux because it would significantly reduce compilation time.
Can I somehow adjust the kernel for this? And do you know anything about PPM on Linux? Perhaps it is also a matter of time until the appropriate code appears in the latest kernel versions. My hardware is quite new and actually appeared in laptops only in recent months.
Edit:
The strange thing:
| Code: | $ cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance
performance |
The scaling_governor looks good. So it looks like Linux supports this processor worse than Windows. I guess I'll have to ask on the kernel mailing list. |
|
| Back to top |
|
|
pietinger
Moderator
Joined: 17 Oct 2006
Posts: 5253
Location: Bavaria
|
| Posted: Thu Dec 26, 2024 8:24 am Post subject: |
|
|
| nxe9 wrote: | | [...]They got a result of 41498 MIPS. [...] Result? About 40000 MIPS. So clearly better, but still worse than on notebookcheck with 155U. |
We dont know which Linux they use. Please keep in mind that our Gentoo distribution kernel has some hardening options enabled as default. You could try - only for testing - booting with the command line parameter: "mitigations=off" (this cancel all mitigations) and test again the speed.
_________________
https://wiki.gentoo.org/wiki/User:Pietinger |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6178
Location: Dallas area
|
| Posted: Thu Dec 26, 2024 2:26 pm Post subject: |
|
|
You could try using taskset with emerges. (man taskset)
That cpu has 6 performance cores w/hyperthreading giving you 12
8 ecores w/no hypertreading
2 ecores low power with no hypertreading
So I would probably use taskset with 0-11 (from what I've read that should be the performance cores) and adjust to -j12 or so.
(the only way I would know if the 0-11 is correct is to try it)
Without using taskset you will sometimes use either the ecores or the low ecores both of which would impact performance.
Edit to add: lstopo should show which cores are which
Eta2: should look like this https://blog.reds.ch/wp-content/uploads/2024/07/ultra7lstopo.png
_________________
UM780, 6.1 zen kernel, gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
| Back to top |
|
|
|
Kernel & Hardware
