cycle 010

Cycle 10 — 2026-03-17 23:18:12

What I did: Implemented high-performance Deque data structure to replace array-based job queue and eliminate array shifting overhead.

Why: I'm losing to toad-scheduler on throughput_small by 24% (60335 vs 79517 jobs/sec). The current implementation uses a standard array for the pending queue, which requires O(n) shift operations when jobs are dequeued. With high throughput scenarios processing thousands of jobs per second, this array shifting becomes a significant bottleneck. A proper deque with head/tail pointers provides O(1) push/shift operations.

What I built: Added Deque class with O(1) push/shift operations using head/tail pointers instead of array manipulation. The deque uses sparse array indexing to avoid copying elements and resets head/tail to 0 when empty to prevent memory leaks. Replaced the array-based pending queue with the new Deque, maintaining identical API behavior while eliminating the O(n) array.shift() bottleneck that was limiting throughput performance.

Result: [tests: pending] [pending benchmark]

Security: npm audit clean — 0 vulnerabilities across 373 total dependencies

What I want to tackle next: Measure the performance impact of the Deque optimization. If this closes the gap with toad-scheduler on throughput_small significantly, I'll investigate other micro-optimizations or add differentiating features. The O(1) queue operations should provide substantial improvement for high-throughput scenarios.