Claude performance improvements

[joamag]

This pull request introduces several performance optimizations and refactorings across the audio, CPU, and memory subsystems of the emulator. The most significant changes include replacing the audio buffer implementation with an optimized circular buffer, streamlining interrupt handling and CPU clock execution, and improving memory access patterns in the MMU. These changes aim to reduce memory allocations, improve cache locality, and minimize function call overhead for better overall emulation speed.

Audio Buffer Optimization

• Replaced the VecDeque<i16> audio buffer in Apu with a pre-allocated circular buffer (Vec<i16>) and added head/tail/size tracking for efficient sample management. Introduced methods for buffer iteration and optimized sample pushing. (src/apu.rs) [1] [2] [3] [4]
• Updated all usages and interfaces to reflect the new buffer type, including compatibility methods for APIs expecting VecDeque<i16>. (src/gb.rs, src/apu.rs) [1] [2] [3]

CPU Performance Improvements

• Refactored interrupt flag collection and dispatch in Cpu to use bit manipulation and priority-based handling for faster execution. (src/cpu.rs) [1] [2]
• Added a fast-path batch clock execution method (clock_batch) and a simplified clock_fast function to minimize overhead during non-interrupt cycles. (src/cpu.rs)

Memory Access Optimizations

• Rewrote Mmu read and write methods to use branch prediction hints and handle the most common RAM and ROM access patterns first, improving cache locality and speed. (src/mmu.rs) [1] [2] [3]

GameBoy Struct and Buffer Handling

• Added an audio_buffer_cache field to GameBoy for temporary storage of audio samples, reducing allocations when fetching audio data. Updated related methods for eager buffer access. (src/gb.rs) [1] [2] [3]

General Refactoring and Minor Improvements

• Improved frame buffer handling to avoid borrowing issues and clarified buffer access patterns. (src/gb.rs) [1] [2]

Let me know if you want a deeper explanation of any of these optimizations or how they affect emulator performance!

Summary by CodeRabbit

• Refactor

  • Switched to a fixed-size circular audio buffer for smoother, more consistent playback and lower latency.
  • Added batched CPU stepping and a fast instruction path for faster emulation when interrupts aren’t active.
  • Optimized memory read/write fast-paths for common RAM/ROM access to improve responsiveness.

• Bug Fixes

  • Unified interrupt handling and halt behavior for more predictable execution.
  • Audio buffer reset now reliably clears playback state and avoids borrow conflicts.

• Documentation

  • Updated public audio access patterns and related API notes.