Runtime PM vs System Sleep

Session information has not yet been published for this event.

Scheduled: Wednesday, September 7, 2011 from 2:50 – 3:40pm in Alexander Valley Ball Room

One Line Summary

There are fundamental differences between runtime PM and system suspend such that system sleep cannot be regarded as "deep idle".


On some systems runtime power management (cpuidle and I/O runtime PM combined) is sufficient to put the system into a low-power state that from the power consumption perspective is not different from a system sleep state. However, while runtime power management seeks to be transparent from the user space’s point of view, it is generally expected that system suspend can happen at any time at the user’s request. This requires system suspend to be carried out in a specific way involving operations that are never performed during transitions initiated by runtime power management. I will explain why system sleep cannot be regarded as a special case of runtime power management in general and why the kernel’s runtime power management frameworks are not generally suitable for implementing system suspend.


PM runtime sleep low-power Linux kernel suspend cpuidle

Presentation Materials



  • Rjw

    Rafael Wysocki

    Faculty of Physics U. Warsaw, SUSE/Novell


    Rafael is the maintainer of the Linux kernel’s core power management subsystem and the suspend-utils package containing hibernate user space tools (known as uswsusp) and the suspend-to-RAM utility called s2ram.

    He works at the Faculty of Physics, University of Warsaw, as a Senior Lecturer. He has been working at the university for over 7 years teaching computer programming and IT and administering computer systems.

    He also works for SUSE Labs as a kernel developer. He has been actively working on the Linux kernel since 2005, mostly on the suspend and hibernate subsystem as well as on power management in general and the ACPI and PCI subsystems.

    Rafael holds a PhD in Theoretical Physics from the University of Warsaw.