Am Mittwoch, 2. September 2020, 07:18:39 CEST schrieb david:
I don't remember what I did on my old i7 to keep
it on performance. One
involved having to push the performance setting to EACH CPU/thread, I
think that's covered on that link somewhere. May have been a script. I
don't remember how I did it, but it stuck between boots. I think I
actually tried three different things, so I really don't know which one
did the trick.
As far as I remember for SysV intit systems the way to make the scaling governor
persistent across reboots was to put a command into /etc/init.d/boot.local or some other
boot init script. The command would vary with distros, whether they used cpufreq-utils or
the newer cpupower suite for this. For cpupower, the command would be
"/usr/bin/cpupower -c all frequency-set -g performance".
Systems using systemd would accordingly need a service file added as
/etc/systemd/system/cpupower.service, and possibly enabled before it would be run at every
boot. I post mine, which I remember having created myself, but I don't remember where
I got it from. Anyway this works well for me:
----
[Unit]
Description=CPU powersave
[Service]
Type=oneshot
ExecStart=/usr/bin/cpupower -c all frequency-set -g performance
[Install]
WantedBy=multi-user.target
----
I haven't done it on my present i9. Laptops aren't really good homes for
the i9 - needs a whole lot more airflow than a laptop can provide.
Running at 900MHz right now and 115F.
Back to HW timers, it's trivial to set them
as explained in the wiki,
but I'm just wondering if they are actually still used?
Don't know about that. Wasn't that something that had to be set in the
kernel at compile time?