Wednesday, April 4, 2012

Free mobile applications can drain batteries two times faster !!!



Recently a study done by Purdue University and Microsoft, found that running free mobile application can drain the battery power two times faster than the regular usage and also reduce the life of the battery. A student called Abhinav Pathak and two of his colleagues designed a unique energy profile tool called "eprof" to analyse the power consumed by some popular Android applications. 



The Apps are Angry Birds, MapQuest, Free Chess, Android browser and New York Times.

The team discovered that around 10-30% of an applications power consumption is some what directly related to its purpose like for example in chess algorithms being calculated in game called Free Chess. While the remaining 70-90% is consumed by the function's that uses wireless, I/O chips that includes Wi-Fi, 3G and GPS etc. In Angry birds around 70% of the energy drain went for downloading & even displaying ads and also for uploading data like location information.

When the processing technology has advanced a lot, the developers have not focused enough on optimising I/O energy that are being consumed by the Applications. Lets take an example most of the Applications have I/O "tail" as it leaves the connections open for a long period after sending or receiving information, this result in wasting the energy. If we tried to understand the data right the "3G tail" in Angry Birds application consumed around 52% of the application energy that is as per the calculations 24% when running and 28%  is taken for the process of closing.


The Free Chess application had a similar breakdown, the Android browser and the New York Times application that we mentioned before spent around 15% on user tracking  while an I/O-heavy app like "MapQuest" proved to be particularly wasteful between 3G and GPS tails. Eprof can also flesh out bugs.The researchers discovered a flaw in Facebook's application that prevented the CPU from entering sleep after the app was closed.

0 comments:

Post a Comment