Skip to contentSkip to conversation

3.10.0: Full dynticks = high load

DISCUSSION Dâniel Fraga 4 participants 8 messages Jul 11, 2013 – Jul 15, 2013

lore
Share & tools
lore
#0 lore raw

I upgraded to kernel 3.10.0 and decided to try the new "Full
dynticks system (tickless)" option, but now the system load is always at 1 or above.

Using the previous "Idle dynticks system (tickless idle)" solves
the problem, so it seems the new Full dynticks code is causing this.

Maybe Ingo Molnar can help to debug this (I put this in Power
Management, but if it should be in other area, please change).

My desktop is:
Asus P8Z68-V Pro/Gen3
Core i7 2700k
8GB RAM

Feel free to suggest how can I debug this to help you. I can try
patches if you want.

Thank you.

Signature · 3 lines
Linux 3.10.0: Unicycling Gorilla
http://www.youtube.com/DanielFragaBR
http://www.libertarios.org.br
#1 lore raw
Quoted 7 lines · Dâniel Fraga

I upgraded to kernel 3.10.0 and decided to try the new "Full
dynticks system (tickless)" option, but now the system load is always at 1 or above.

Using the previous "Idle dynticks system (tickless idle)" solves
the problem, so it seems the new Full dynticks code is causing this.

This describes exactly what I have encountered, and "tickless idle" fixed it for me, too. So I'll second that.

#2 lore raw
Quoted 2 lines · On Fri, 12 Jul 2013 08:52:27 +0200 Heinz Diehl <[email protected]> wrote:

This describes exactly what I have encountered, and "tickless idle" fixed it for me, too. So I'll second that.

Thanks for the confirmation. I hope Ingo Molnar fixes it.

Signature · 3 lines
Linux 3.10.0: Unicycling Gorilla
http://www.youtube.com/DanielFragaBR
http://www.libertarios.org.br
#3 lore raw

(add cc)

Quoted 8 lines · Dâniel Fraga
Quoted 2 lines · On Fri, 12 Jul 2013 08:52:27 +0200 Heinz Diehl <[email protected]> wrote:

This describes exactly what I have encountered, and "tickless idle" fixed it for me, too. So I'll second that.

Thanks for the confirmation. I hope Ingo Molnar fixes it.

#4 lore raw
Quoted 13 lines · Mike Galbraith

(add cc)

Quoted 8 lines · Dâniel Fraga
Quoted 2 lines · On Fri, 12 Jul 2013 08:52:27 +0200 Heinz Diehl <[email protected]> wrote:

This describes exactly what I have encountered, and "tickless idle" fixed it for me, too. So I'll second that.

Thanks for the confirmation. I hope Ingo Molnar fixes it.

Hi,

Sorry I'm missing the description of the issue. Could you please repaste it?

Thanks!

#5 lore raw
Quoted 6 lines · On Mon, 15 Jul 2013 16:25:59 +0200 Frederic Weisbecker <[email protected]>…

Hi,

Sorry I'm missing the description of the issue. Could you please repaste it?

Thanks!

The problem is that when we use the new Full dynticks option
form kernel 3.10.0, the load will go high, always above 1. Bug?

Signature · 3 lines
Linux 3.10.0: Unicycling Gorilla
http://www.youtube.com/DanielFragaBR
http://www.libertarios.org.br
#6 lore raw
Quoted 5 lines · Dâniel Fraga
Quoted 2 lines · On Mon, 15 Jul 2013 16:25:59 +0200 Frederic Weisbecker <[email protected]>…

Hi,

Quoted 7 lines
Quoted 4 lines

Sorry I'm missing the description of the issue. Could you please repaste it?

Thanks!

The problem is that when we use the new Full dynticks option
form kernel 3.10.0, the load will go high, always above 1. Bug?

You mean this is the result that the "uptime" command gives you? And your system is doing something? Or is it mostly idle?

#7 lore raw
Quoted 6 lines · On Mon, 15 Jul 2013 17:01:35 +0200 Frederic Weisbecker <[email protected]>…
Quoted 2 lines · Dâniel Fraga

The problem is that when we use the new Full dynticks option
form kernel 3.10.0, the load will go high, always above 1. Bug?

You mean this is the result that the "uptime" command gives you? And your system is doing something? Or is it mostly idle?

It is idle, but the load remains high all the time. Going back
to the "Idle dynticks system (tickless idle)" solves the problem.

So the issue is with the new Full dynticks.

Signature · 3 lines
Linux 3.10.0: Unicycling Gorilla
http://www.youtube.com/DanielFragaBR
http://www.libertarios.org.br