Is Linux Power Management Getting Better Or Worse?

We began with the simple test of letting the ThinkPad notebooks idle within the GNOME desktop followed by continuing the idle process but with the LVDS panels signaled off via DPMS (this was done via the Phoronix Test Suite's battery-power-usage test profile). Interestingly, neither Fedora 13 nor Fedora 14 Alpha were the most power efficient releases with this most basic test except for Fedora 14 Alpha on the vintage ThinkPad R52. With the R52 testing we could not run Fedora 11 or 12 due to installation issues, but for this Pentium M notebook its power rate was lower by nearly 20% when running this newest development release. For both the T60 and T61 with its newer Intel dual-core hardware, the power consumption rate increased in both Fedora 13 and Fedora 14 Alpha after both experienced a nice improvement in Fedora 12.

Next, we looked at the power consumption rate while the notebooks had their CPUs taxed with the OpenMP-powered GraphicsMagick. Fedora 14 Alpha again did the best when it was running on the ThinkPad R52, but with the Core Duo and Core 2 Duo notebooks the battery life was shorter than with Fedora 13. These decreases in battery life may be due in part though due to debugging options built with the current Fedora 14 development packages that may be keeping up the notebooks up more frequently and taxed than they are in released versions of Red Hat's playground.

For those wondering whether the performance-per-Watt though is changing between releases, with GraphicsMagick the performance really has not changed greatly over the past four releases (GCC on Fedora 10 had not worked with the tested version of GraphicsMagick).