Eventually "Schedutil" Could Replace Linux's Existing CPU Scaling Governors

[phoronix]

Phoronix: Eventually "Schedutil" Could Replace Linux's Existing CPU Scaling Governors

The Schedutil CPU frequency scaling governor has been around for a few years and has gotten better over time but in our own tests we still find it frequently not being as competitive to the "performance" governor and others. However, in the future Schedutil might become the default and perhaps only governor...

http://www.phoronix.com/scan.php?page=news_item&px=CPUFreq-Schedutil-Future-Maybe

[Linuxxx]

Honestly, 'schedutil' sounds awesome in theory!
Here's what a SUSE Labs engineer has to say on the matter:

Schedutil is a cpufreq governor that uses scheduler utilization data (the PELT number) to select CPUs' operating frequencies. The PELT number is per-process (as opposed to being per-core), so its value isn't lost when the process migrates across CPUs -- this is beneficial for making decisions about how fast the CPU has to run. Schedutil is fundamentally different from the current best-in-class governors, that read per-core utilization values (such as intel_pstate on x86) or delegate the matter of frequency selection entirely to the platform (HWP on x86, CPPC on ARM). We'll review schedutil's recent developments and present experimental results of its performance in a few practical scenarios.

Giovanni Gherdovich

[amdtesterman]

the question is if there is any similar situation with intel and them intel_pstate.. why not to do the same in amd cpu more or less? or if it is the same, why kernel mantainers don't want it and yes the intel ones..

[Linuxxx]

Even though "schedutil" still seems rather buggy, it's already successfully being used by Android to provide improved latency & smoothness.
Also, here's what a fellow Linux-CK user had to say about it:

primusrun and nvidia-xrun with intel_cpufreq schedutil makes all Valve games I've played open several seconds faster and leaves me with unbelievably low mouse latency on an Optimus system compared to mainline and Windows.

[DarkFoss]

Schedutil is great if you enjoy stuttering in games.

[Peter Fodrek]

. Peter Zijlstra was critiquing several points of the proposed AMD CPPC code that AMD developers have been working on for Collaborative Processor Performance Control support with their new Zen 2 processors.Zijlstra rejected various elements of the AMD code proposal, including the calling it a "huge mistake

I think it is new DAL/DC soi it will be accepted in the future
AMD's Massive New DAL Display Driver Is Facing An Uphill Battle
12 February 2016 at 08:02 AM EST
We weren't the only ones surprised by the massive size of the new AMDGPU DAL display driver. Weighing in at 93k lines of cod

https://www.phoronix.com/scan.php?page=news_item&px=Massive-DAL-Driver-Woes

It Looks Like AMDGPU DC (DAL) Will Not Be Accepted In The Linux Kernel 8 December 2016

https://www.phoronix.com/scan.php?page=news_item&px=AMDGPU-DC-DRM-No

AMDGPU DC Code Lands For Linux 4.15 Kernel on 17 November 2017
Linus Torvalds has accepted the AMDGPU DC display code pull request for the Linux 4.15 kernel. AMD Linux users can now rejoice!

https://www.phoronix.com/scan.php?page=news_item&px=AMDGPU-DC-Accepted

[timofonic]

I think it is new DAL/DC soi it will be accepted in the future
AMD's Massive New DAL Display Driver Is Facing An Uphill Battle
12 February 2016 at 08:02 AM EST
We weren't the only ones surprised by the massive size of the new AMDGPU DAL display driver. Weighing in at 93k lines of cod

https://www.phoronix.com/scan.php?page=news_item&px=Massive-DAL-Driver-Woes

It Looks Like AMDGPU DC (DAL) Will Not Be Accepted In The Linux Kernel 8 December 2016

https://www.phoronix.com/scan.php?page=news_item&px=AMDGPU-DC-DRM-No

AMDGPU DC Code Lands For Linux 4.15 Kernel on 17 November 2017
Linus Torvalds has accepted the AMDGPU DC display code pull request for the Linux 4.15 kernel. AMD Linux users can now rejoice!

https://www.phoronix.com/scan.php?page=news_item&px=AMDGPU-DC-Accepted

What are you saying. I cannot understand you. Please elaborate, you don't need to do broken quoting of everything.

[Venemo]

Sounds to me like AMD can still rework their CPPC patches to work inside schedutil instead of being a separate governor.

[amdtesterman]

The question is why not to say the same to intel with intel pstate. Is my only one question. Because I dude amd code is so different from intel ones but the old age that have intel ones

[Peter Fodrek]

What are you saying. I cannot understand you. Please elaborate, you don't need to do broken quoting of everything.

Only that AMD CPPC will be same as AMD DAL/DC Firstly rejected and then a accepted