Baking Android with different memory allocaters
Submitted by satish patel (@satishp7) on Friday, 9 September 2016
Crisp talk (15 minutes)
Android has not just limited its self to mobile space but has started making its brand to all other consumer domains. And it raises a need of customizing/optimizing Android for space & time as per individual needs. This presentation will highlight on different malloc allocaters and its chemistry with Android at runtime. Malloc is one of the crucial component for speed, because if malloc is not fast enough, application writers are inclined to write their own custom free lists on top of malloc. This can lead to extra complexity, and more memory usage unless the application writer is very careful to appropriately size the free lists and scavenge idle objects out of the free list.
The presentation will talk about efforts made to integrate various memory allocaters – nedmalloc, tlsf, musl, tcmalloc etc. to Android libc and present space & time complexity results for baked Android. Talk will broadly focus on integration challenges, memory benchmarks, fragmentation results, comparison with master AOSP and finally conclusion on which allocater can be best choice for use case.
- Why Android
- Memory allocaters
Satish Patel is an Android engineer at Linaro. He has extensive experience across Open Source, Linux drivers, Android porting & internals, EMV L1, Social Network Service Framework and other cutting edge technologies in the embedded domain. He is secretary of Bangalore Chapter of the IEEE Consumer Electronics Society.At Linaro, Satish works on Android customization/optimization. Before joining Linaro he has worked with Texas Instruments, Samsung and eInfochips.