The Linaro team’s Android ROM and tweaks that I talked about yesterday certainly promised some impressive results, so when I found a reasonable-looking source for a Galaxy Nexus ROM that supposedly has those features, curiosity got the best of me. Coming from the developers of the Android Open Kang Project, or AOKP ROM, the build is basically the latest version of AOKP rebuilt to include most of Linaro’s tweaks and features. Obviously, the software is still in development and won’t be a viable daily driver, but I still decided to give it a try on my Galaxy Nexus to see if Linaro can live up to all the hype.
Since I didn’t really want to start with a vanilla installation for testing, I decided to install Linaro over my current CyanogenMod installation without wiping any data. In addition to being easier than installing from scratch, this also allowed me to compare the build of CyanogenMod that I have been using complete with all the applications and potential bloatware. Because a fresh installation always feels faster than one that has been in use, I figured this would be a good way to test the real performance of Linaro. Although I don’t have any video because screencast apps don’t yet support Ice Cream Sandwich, you will have to take my word for it that Linaro is fast, even in the alpha state. Scrolling on homescreens works as smooth as butter; the app drawer, the browser, and multitouch gestures work great, as well. This is, of course, subjective, but it also seems like applications open much faster than on CyanogenMod, and task switching works much more smoothly.
I did run a few benchmarks, and interestingly Linaro did not do quite as well as I expected. In fact, it scored lower in Quadrant than a stock Galaxy Nexus, and fairly average on things like Linpack. Additionally, Bernhard Rosenkranzer, the man running the benchmarks in the previous post about Linaro, has said on Google+ that the massive improvement was likely something of a fluke. Yes, Linaro will improve Android’s performance quite a bit, but perhaps not as much as was suggested by the initial benchmark.
Despite lackluster real world benchmarks, I still think that Linaro is a big improvement over stock or even CyanogenMod Android. In fact, ROM makers often champion benchmarks as the ultimate measure of ROM performance, when realistically what matters is real world performance. So, while Linaro may no longer be busting the benchmarks like some expected, in my actual use of the ROM I have found that it does seem to have a significant impact on the difficult to measure real world performance of a ROM. However, you won’t be able to really get the full picture until you try Linaro for yourself, which you can only do right now if you have a Galaxy Nexus. If you do, then just follow the source link to get downloads for your device, but if not, I would wait for the soon to come nightly builds of CyanogenMod that will include the Linaro tweaks. Even if they won’t impress people with benchmarks, the actual in use performance is what matters, and from what I can tell Linaro is a fairly significant improvement over the current Android software in that area.