Performance Tuning

fink. is fast out of the box, but power users can tweak the engine flags to squeeze every drop of performance out of their hardware.

The Rust Core

Unlike Chromium, which relies on a massive C++ codebase, fink.'s rendering pipeline and Javascript bridge are written in memory-safe Rust. This allows us to aggressively multithread DOM layout without fearing race conditions.

Memory Purging

By default, fink. aggressively unloads inactive tabs from RAM after 15 minutes of inactivity. The DOM state is serialized to NVMe storage, meaning restoring the tab takes less than 50ms, but frees up gigabytes of RAM.

You can modify this timeout in your config:

json

{
  "performance": {
    "tabSuspendTimeoutMs": 900000,
    "serializeToDisk": true
  }
}

Hardware Acceleration Flags

fink. utilizes WebGPU by default for all 2D canvas and CSS animations, offloading work from the main thread.

If you are on an older integrated GPU (e.g., Intel HD 4000 series), forcing WebGPU might cause visual artifacting. You can fallback to OpenGL by setting "forceOpenGL": true in your engine flags.

Local LLM Constraints

The built-in AI assistant runs locally on your machine. By default, it allocates 4GB of RAM for the quantized model weights. If you notice system-wide slowdowns when chatting with the AI, you can downgrade to the smaller 2GB fink-mini model via the Command Palette.

PreviousLayouts & Sidebars NextPrivacy Options