TechPrufer
Get started
Meta
Meta logoMeta · Feature trail

Primitive Data Structures for Android

2 posts · Jun 2016

What this trail covers

Meta's ongoing development in mobile engineering includes the creation of specialized primitive data structures for Android to optimize performance and memory usage. This involves implementing custom Set implementations like LongArraySet and IntArraySet, which leverage Android's SparseArrayCompat and LongSparseArray for efficient storage and retrieval of primitive types. These structures are designed to offer performance benefits over traditional Java collections for specific use cases, particularly by reducing object allocations and mitigating garbage collection pauses.

All Meta trails
  1. Primitive sets for AndroidThis post introduces two new primitive Set implementations for Android: LongArraySet and IntArraySet. These classes are built upon Android's LongSparseArray and SparseArrayCompat respectively, providing efficient storage and retrieval of primitive long and int values. The implementations include optimizations for adding, removing, and checking for the existence of elements, with a note on their performance characteristics compared to traditional HashSets. The goal is to offer more memory-efficient and potentially faster alternatives for specific Android development scenarios.
  2. Memory optimization for feeds on AndroidThis post details the creation and evaluation of `LongArraySet` and `IntArraySet` for Android. These custom data structures were developed to address excessive object allocations caused by Java's autoboxing feature when using `HashSet<Long>` and `HashSet<Integer>` for storing primitive `long` and `int` values. By leveraging `LongSparseArray` internally, `LongArraySet` avoids creating `Long` objects, significantly reducing memory usage and garbage collection overhead. Performance benchmarks showed improved `contains` and `remove` operations for moderate data sizes, with a trade-off of slower `add` operations for very large datasets. The implementation resulted in a ~30% reduction in allocations compared to `HashSet` in specific scenarios.