Using 64-bit Linux Software To Reduce Power Consumption?

With Ubuntu 16.04 LTS due for release next month, I found it time to do my occasional 32-bit vs. 64-bit Linux OS comparison for showcasing the performance difference of using the 64-bit software on supported hardware. This time around, at a premium member's request, I also did some power consumption monitoring between the 32-bit and 64-bit builds of Ubuntu 16.04 on an Intel laptop/ultrabook.

The 32-bit vs. 64-bit Ubuntu 16.04 LTS raw performance benchmarks will be published tomorrow on Phoronix. For those interested in the power consumption difference, those results are what can be seen today. The last time I did any power monitoring when comparing 32-bit and 64-bit of the same software/distribution, I didn't recall seeing any real difference. With this 16.04 LTS testing on a Haswell notebook, it was a bit more noticeable.

An Intel UX301LAA ultrabook with Intel Core i7 4558U CPU, 8GB of RAM, and dual 128GB SanDisk SSDs were used for testing. The same Ubuntu 16.04 daily ISO build was cleanly installed each time on this ultrabook with its Linux 4.4.0-11-generic kernel, Unity 7.4.0, Mesa 11.2, GCC 5.3.1, and P-State powersave configuration.

Over the course of running a number of tests, the 64-bit Ubuntu had an average power use of 30.2 Watts while the 32-bit Ubuntu had an average power draw of 31.9 Watts. The lowest power consumption was also seen with the 64-bit build at 8.5 Watts vs. 9.4 Watts. The 32-bit Ubuntu also peaked higher at 54.3 Watts compared to 49.7 Watts.

The power consumption was clearly lower when using the 64-bit version of Ubuntu 16.04 over the i686 build.

It's with the Xonotic gaming test where the power use was fluctuating much more:

Stay tuned for more results. Let me know if you're interested in more 32-bit vs. 64-bit Linux power testing from other hardware to complement tomorrow's 32-bit vs. 64-bit raw performance benchmarks; if you are part of premium, I try to honor all those requests.