How Flutter Manages Memory: Dart Runtime, Image & FlutterView Insights

Dart Runtime Overview

Flutter framework is built with Dart, so each app includes a Dart VM. The VM compiles Dart to native bytecode via AOT, eliminating interpretation and improving performance. The article focuses on Dart VM memory allocation and garbage collection.

Dart isolates have separate heaps and stacks, each with its own GC, communicating via message channels. This isolation prevents data races; the entire Flutter widget rendering runs in a single isolate.

Exploring Image Memory

Flutter provides an Image widget. The article investigates whether Android ImageView uses Java heap or native memory by adding images incrementally and profiling with Android Profiler on Android 6.0, 7.0, and 8.0.

Results show Java heap growth on Android 6.0/7.0 and native memory growth on Android 8.0, indicating ImageView uses Java memory on older versions and native memory on newer ones.

Flutter Image uses neither Java nor native memory but Graphics memory, avoiding OOM risks and leveraging native memory.

Flutter's ImageCache caches images by count rather than precise memory size.

FlutterView Memory Exploration

When integrating Flutter into existing apps, developers must decide whether to create a new FlutterView per Activity or reuse an existing one. Reusing requires attaching the FlutterNativeView to each Activity, which can introduce memory‑leak risks if not detached properly.

Tests with empty pages show that not reusing a FlutterView adds about 0.73 MB native memory per page, while reusing adds about 0.65 MB, indicating native memory savings.

Reusing a single FlutterView can cause visual glitches during Activity switches; a possible mitigation is to capture a screenshot to mask the transition.

A hybrid approach using a FlutterViewProvider that maintains multiple reusable FlutterViews balances reuse benefits with stability.

Pre‑warming the first frame of a FlutterView in a 1‑pixel window can reduce startup latency.

The article concludes with practical insights from the Xianyu team on Flutter memory management and encourages further experimentation.