Power consumption results in full for a range of mobile devices when running the latest development snapshot of Ubuntu 12.04 LTS compared to earlier Ubuntu Linux releases will be published on Phoronix next week. In this article are just the results from an old and newer laptop to illustrate Ubuntu 12.04 being in great power shape and at its best of the latest releases, while for older hardware the Precise Pangolin can be very power hungry.
The Ubuntu 12.04 LTS power consumption testing follows this week's Ubuntu 12.04 LTS boot performance results for six different notebooks/netbooks and the results going as far back as Ubuntu 6.06.1 LTS "Dapper Drake" from six years ago. Like the to-be-published power results, the Bootchart results were also mixed. For the newer hardware, the boot speed for Ubuntu 12.04 LTS tended to improve for some while on the older hardware it's seen faster days out of the older Ubuntu releases.
PCI Express ASPM power regression I pointed out last year and then was finally fixed in a proper manner during the 12.04 cycle. There's also been other power optimizations that Canonical engineers and other upstream developers have been working on to try to improve the power consumption of Ubuntu Linux, but this largely benefits newer hardware.
The power consumption results of Ubuntu 12.04 LTS for a spectrum of mobile x86 hardware will be published next week. Besides looking at the average battery power consumption when idling, there will also be power consumption results when under full load of the CPU, GPU, etc. The power results largely go along with my comments made earlier this week where If You Have Old Hardware, Think Twice About Ubuntu 12.04 LTS. If you have older hardware (4~5+ years old, such as anything with a CPU that's 32-bit only), you're likely better off sticking with Ubuntu 10.04 LTS, similarly-aged distributions, or a lightweight Linux distribution rather than trying to run Ubuntu 12.04 LTS. The overall performance for this older hardware is largely in a free-fall.
Testing was facilitated in an automated manner via the Phoronix Test Suite. ACPI battery (and various USB-based power meter) monitoring can be done simply having the MONITOR environment variable set to sys.power (or all) when running a test (e.g. MONITOR=all phoronix-test-suite benchmark idle c-ray).