⚡ Mobile Performance Pass (April 2026)
This performance pass targeted long-session smoothness on mobile. The most visible player complaint was not a total crash; it was micro-stutter after two to three continuous rounds in high-motion games. We ran repeatable tests on iPhone 13 Safari, Pixel 7 Chrome, and Windows 11 Chrome, then compared frame stability, memory behavior, and restart speed for Neon Swarm, Escape Rush, and Neon Flap.
Test Environment
- iPhone 13 (iOS 17.4, Safari, low power mode off)
- Pixel 7 (Android 14, Chrome 135, battery saver off)
- Windows 11 desktop (Chrome 135, 60Hz panel)
- Network profile: Wi-Fi 5, first load uncached and second load cached
Measured Results
| Metric | Before | After | Notes |
|---|---|---|---|
| Median frame time (mobile action scenes) | 19.8ms | 16.9ms | Reduced frame jitter during particle-heavy moments |
| P95 frame time spike | 42ms | 27ms | Lower worst-case stutter during enemy waves |
| Restart-to-play ready | 1.24s | 0.68s | State reset and overlay transition cleanup |
| 30-minute memory growth (Pixel 7) | +86MB | +31MB | Audio/particle references released earlier |
What Changed
The first fix was lifecycle discipline. We tightened object cleanup after each game-over state, especially around particle arrays and temporary UI nodes that were previously retained longer than needed. The second fix was draw-call pressure: visual effects that looked similar but consumed extra paint budget were replaced with simpler variants in peak scenes. The third fix was restart path trimming so a new run does not wait on non-critical transitions.
These changes are intentionally conservative. We did not remove core visual identity or reduce gameplay speed; we only reduced work that did not directly improve player decisions. That keeps the "feel" intact while reducing dropped frames in edge cases.
Player Impact
- Less visible hitching during high-enemy moments in action games.
- Faster retry loop, which improves learning pace and reduces frustration.
- More stable behavior on mid-range phones over repeated sessions.
- Lower chance of delayed input perception caused by frame spikes.
Next Validation Step
The next pass will focus on heat and battery impact during 20-minute sessions. We plan to publish per-device battery drain ranges and note any compromises if effects need additional tuning. This helps keep our "quick, smooth, no-install" promise measurable rather than purely descriptive.