Linux 6.9 Has A Big Rework To CPU Timers - Some Power/Performance Benefits

[phoronix]

Phoronix: Linux 6.9 Has A Big Rework To CPU Timers - Some Power/Performance Benefits

The Linux 6.9 kernel has a big rework to the CPU timer code that has been years in the making and has some power and performance benefits...

https://www.phoronix.com/news/Linux-6.9-Timers-Rework

[aviallon]

I would be interested to see how the worst-case latency of network irqs is affected by that.

[Kjell]

The amount of required experience and knowledgeable to pull of such refactoring is astonishing.

Really impressive work!

[uid313]

I don't know what this is about, but it is 2024 and it still occurs that under heavy load my system becomes unresponsive to keyboard presses and the mouse cursor not moving when I move the mouse.

[and.elf]

I don't know what this is about, but it is 2024 and it still occurs that under heavy load my system becomes unresponsive to keyboard presses and the mouse cursor not moving when I move the mouse.

Maybe limit process resources using containers or cgroups?
The only time this has happened to me was on a single core cpu and intentional fork-bomb

[ATLief]

I don't know what this is about, but it is 2024 and it still occurs that under heavy load my system becomes unresponsive to keyboard presses and the mouse cursor not moving when I move the mouse.

If you (or and.elf) are looking for easy drop-in resource management, you may be interested in Simple Slices. It's not done yet, but totally usable.

[EphemeralEft]

Does that mean with 12 cores, each of 8 cores will be handling 2.7x less timer interrupts than each of the other 4 cores?

If so, is that basically like setting 8 cores to tickless?

[caligula]

The amount of required experience and knowledgeable to pull of such refactoring is astonishing.

Really impressive work!

Yes, there are standard "full-stack" MERN Mongo/Express/React/Node developers and kernel developers. On the opposite ends of the bell curve.