Is Intel's PowerTOP Utility Still Beneficial In 2016 On Ubuntu 16.04 To Save Power?
Intel OTC's PowerTOP utility has been around for nearly a decade for making it easy to carry out power optimization tweaks on Intel Linux systems. However, is this program still useful or are modern Linux distributions and upstream code now better optimized by default for delivering an ideal power-savings experience? As it's been a while since the last time I tried PowerTOP, I fired it up today on an Intel Haswell ultrabook running a development snapshot of Ubuntu 16.04.
It's been a while since I last tried PowerTOP and with Ubuntu 16.04 being around the corner, I decided to try it out an ASUS UX301LAA Zenbook for seeing the power impact in 2016. For years it's been the easy way of extending your battery life on Linux.
The test system I tried it on had a Core i7 4558U processor with Haswell-ULT graphics, 8GB of RAM, dual 128GB SanDisk SSDs, and a 2560 x 1440 display. The Ubuntu 16.04 snapshot I was using had the Linux 4.4.0-11-generic kernel, Mesa 11.1.2, GCC 5.3.1, and an EXT4 file-system.
I did a before and after test -- one a stock Ubuntu 16.04 experience and then the other when enabling all of the PowerTOP tunables for lowering the power consumption.
Utilizing PowerTOP led this Intel ultrabook to lower its average power use when idling from 12.9 Watts to 11 Watts. The minimum power draw was also much better at 12.7 vs. 10.8 Watts and the peak power use was also lower thanks to the Intel optimizations. Lowering the average power use by 15% when idling thanks to spending two minutes with PowerTOP is quite a win.
Under load, the average power use was slightly better at 39.1 Watts vs. 40.1 Watts. All of the load details can be found via this OpenBenchmarking.org result file. These tests were done via the Phoronix Test Suite while the device was on battery power and the MONITOR=sys.power environment variable set.
Long story short, PowerTOP is still proving to be useful in 2016 on modern Linux distributions with still relatively new Intel mobile hardware for extending the battery life. If you're interested in seeing PowerTOP results from more Linux laptops/ultrabooks, let me know via Twitter or the forums and if there's enough interest (particularly from the premium party), I'll be happy to run some more tests for reference purposes.
It's been a while since I last tried PowerTOP and with Ubuntu 16.04 being around the corner, I decided to try it out an ASUS UX301LAA Zenbook for seeing the power impact in 2016. For years it's been the easy way of extending your battery life on Linux.
The test system I tried it on had a Core i7 4558U processor with Haswell-ULT graphics, 8GB of RAM, dual 128GB SanDisk SSDs, and a 2560 x 1440 display. The Ubuntu 16.04 snapshot I was using had the Linux 4.4.0-11-generic kernel, Mesa 11.1.2, GCC 5.3.1, and an EXT4 file-system.
I did a before and after test -- one a stock Ubuntu 16.04 experience and then the other when enabling all of the PowerTOP tunables for lowering the power consumption.
Utilizing PowerTOP led this Intel ultrabook to lower its average power use when idling from 12.9 Watts to 11 Watts. The minimum power draw was also much better at 12.7 vs. 10.8 Watts and the peak power use was also lower thanks to the Intel optimizations. Lowering the average power use by 15% when idling thanks to spending two minutes with PowerTOP is quite a win.
Under load, the average power use was slightly better at 39.1 Watts vs. 40.1 Watts. All of the load details can be found via this OpenBenchmarking.org result file. These tests were done via the Phoronix Test Suite while the device was on battery power and the MONITOR=sys.power environment variable set.
Long story short, PowerTOP is still proving to be useful in 2016 on modern Linux distributions with still relatively new Intel mobile hardware for extending the battery life. If you're interested in seeing PowerTOP results from more Linux laptops/ultrabooks, let me know via Twitter or the forums and if there's enough interest (particularly from the premium party), I'll be happy to run some more tests for reference purposes.
17 Comments