You are not logged in.
Hello,
I'm trying to optimize power management on my laptop (a Thinkpad Yoga with an Intel core i7 Haswell processor), but I discovered a strange behavior involving PC7 state after suspend-to-disk. I'm using the lastest Arch Linux kernel 3.13, but I also had this problem a few days ago with 3.12.
I've added the following parameters to the kernel command line :
pcie_aspm=force i915.i915_enable_rc6=3 i915.i915_enable_fbc=1 i915.lvds_downclock=1and I set the PCIE policy to powersave at boot time with :
echo powersave > /sys/module/pcie_aspm/parameters/policyWith this configuration, the processor goes up to the package C7 state (PC7) when idling (which is a state with a very low power consumption, possibly the lowest power consumption possible in "normal" usage). This can be verified with the "turbostat" software (%pc7 below) :
cor CPU %c0 GHz TSC SMI %c1 %c3 %c6 %c7 CTMP PTMP %pc2 %pc3 %pc6 %pc7 %pc8 %pc9 %pc10 Pkg_W Cor_W GFX_W
0.60 2.95 2.39 0 0.93 0.04 0.03 98.40 37 37 23.95 0.94 7.03 64.73 0.00 0.00 0.00 1.30 0.07 0.00
0 0 0.75 2.96 2.39 0 0.53 0.07 0.04 98.61 37 37 23.95 0.94 7.03 64.73 0.00 0.00 0.00 1.30 0.07 0.00
0 1 0.13 2.77 2.39 0 1.15
1 2 0.51 2.94 2.39 0 1.27 0.01 0.01 98.19 37
1 3 1.01 2.98 2.39 0 0.77However, after a suspend-to-disk and a resume, the PC7 state is not longer used, even when the computer is completely idling :
cor CPU %c0 GHz TSC SMI %c1 %c3 %c6 %c7 CTMP PTMP %pc2 %pc3 %pc6 %pc7 %pc8 %pc9 %pc10 Pkg_W Cor_W GFX_W
0.72 1.80 2.39 0 1.42 0.18 0.10 97.58 41 41 22.03 73.25 0.00 0.00 0.00 0.00 0.00 2.05 0.04 0.00
0 0 1.96 1.80 2.39 0 1.08 0.24 0.13 96.60 41 41 22.03 73.25 0.00 0.00 0.00 0.00 0.00 2.05 0.04 0.00
0 1 0.02 1.80 2.39 0 3.01
1 2 0.80 1.80 2.39 0 0.45 0.11 0.08 98.57 40
1 3 0.12 1.80 2.39 0 1.13Additionally, changing /sys/module/pcie_aspm/parameters/policy has no effect.
Finally, if I reboot the computer, or if I suspend-to-RAM (even after a suspend-to-disk), PC7 state is activated again.
Do you have an idea of what is wrong ? As far as I know, suspend-to-disk is close to a full reboot, with the RAM saved to the disk, so how the results can differ after reboot and suspend-to-disk ?
Offline
However, after a suspend-to-disk and a resume, the PC7 state is not longer used, even when the computer is completely idling :
[...]
Finally, if I reboot the computer, or if I suspend-to-RAM (even after a suspend-to-disk), PC7 state is activated again.
I dont understand your second sentence... Did you somehow manage to activate it again or not?
In case you got a Haswell CPU please check out this thread as it addresses a similar issue: Haswell-System doesn't go into powerstates lower than pc3
Offline