PR #1651: Optimizing Queue Pop Performance

Continuing my work on Quantinuum’s Guppy: I improved the performance of the standard library’s Queue data structure.

Guppy’s std.collections.Queue provides a fixed-capacity queue for quantum and classical programs. The original implementation used a linear-time algorithm for pop operations — meaning performance degraded as the queue grew larger.

In PR #1651, I replaced this with a circular-buffer implementation that delivers constant-time performance for push, pop, and peek operations. The key insight: track the start and end positions in the buffer using modular arithmetic, so elements wrap around when reaching the end.

The implementation required careful handling of edge cases. A circular buffer with only two pointers can’t distinguish between completely full and completely empty states. After exploring this approach with reviewer feedback, the final solution uses a dedicated size property to track the queue length — ensuring the queue can store its full declared capacity while maintaining O(1) operations.

The performance improvement is significant: benchmark tests showed the push-pop benchmark improved from ~623ms to ~426ms — roughly 32% faster.

Performance optimizations like this matter because they let developers work with larger data structures without worrying about hidden performance costs. It’s satisfying when a well-known algorithmic pattern solves a real-world bottleneck.

Thanks to the Guppy contributors at Quantinuum for their collaboration and thoughtful reviews.

For more technical details, refer to PR #1651.