Google Android 14 reduces background activities to improve performance and extend battery life
CTOnews.com, Oct. 25, in the latest # TheAndroidShow event, Google's vice president of Android Engineering Dave Burke said that Android 14 has changed the workload balancing mechanism of processors and memory, which can not only improve device performance, but also improve battery life performance.
Android 14 will restrict broadcasting (broadcasts) to save resources. CTOnews.com Note: broadcasting is the communication mechanism between the system and the application when critical events occur, such as when battery power drops and Internet connections switch from Wi-Fi to cellular networks.
Mr. Burke said on the show that
Android 14 once again reduced CPU activity for background applications, thus reducing the frequency of cold starts of applications by 30 per cent.
During a cold start, the code page must be read from flash memory and transferred to memory before execution, so the user can clearly sense a 30% reduction.
There are many challenges in reducing cold startup, and we naturally want to increase the number of caching processes (that is, processes in RAM) so that more applications can be hot-started.
But the key here is how to balance. If there are too many cached processes and they are not actually frozen, then these processes will consume processor resources and consume battery power.
Through continuous testing and running-in, the team can reduce CPU footprint by 50% while the application background is running. We achieve this through a number of internal projects, appropriately freezing threads and reducing transmission.
Android 14 also optimizes application-related code to ensure that the application footprint is reduced without compromising performance.
Cold start means that the application starts from scratch. This means that the system process does not create the application process until it is cold started. When cold start occurs, the application starts for the first time after the device starts or after the system terminates the application.
Mr Burke said:
By collapsing the return statement and performing various optimizations in the compiler, we reduced the code size by 9%. If the device has applications that have been optimized by 500MB-1GB, the 50MB-100MB footprint can be reduced compared to before optimization.
