# Measurements — 2026-11-21 (simulator)
Every number on this page comes from the iOS Simulator on a Mac, not from an iPhone. Simulator frame pacing is capped at 60 Hz and runs on desktop hardware, so these figures show that nothing in the games is pathologically slow; they say nothing reliable about a real device, its battery or its temperature.
# Environment
| Item |
Value |
| Host |
Apple M5 Max, macOS 27.0 (26A5425a) |
| Xcode |
27.2 (27B52) |
| Simulator |
"ORIXArcade iPhone" — iPhone 17 Pro, iOS 26.5 (23F77) |
| Build |
0.1.0 (1), Debug configuration, audio muted (-ORIXMute YES) |
| Input |
the games' built-in bots (-ORIXAutoplay YES); frame times logged by -ORIXPerfLog YES |
| Reproduce |
Scripts/build.sh build && Scripts/screenshots.sh (logs land in build/perf/, copied to docs/metrics/raw/) |
# Frame pacing (target: 60 fps, i.e. 16.67 ms)
| Scene |
Frames |
Seconds |
Average |
p95 |
p99 |
Worst frame |
Average fps |
| Traffic Mayhem — Four Corners, Rush |
4,282 |
72.3 |
16.88 ms |
16.67 ms |
16.67 ms |
172.1 ms |
59.2 |
| Traffic Mayhem — Main Street, Rush |
4,315 |
72.3 |
16.76 ms |
16.67 ms |
16.67 ms |
140.8 ms |
59.7 |
| Traffic Mayhem — Twin Junction, Rush |
4,318 |
72.3 |
16.75 ms |
16.67 ms |
16.67 ms |
81.1 ms |
59.7 |
| Domino City — level 15 (longest log) |
7,615 |
128.6 |
16.89 ms |
16.67 ms |
16.67 ms |
189.7 ms |
59.2 |
| Domino City — levels 1–14 (4–8 s each) |
239–481 |
4.0–8.1 |
16.65–17.09 ms |
16.67 ms |
16.67–33.33 ms |
16.7–139.3 ms |
58.5–60.0 |
| Dig to Core — Core Mantle start, maxed drill |
482 |
8.0 |
16.67 ms |
16.67 ms |
16.67 ms |
20.0 ms |
60.0 |
| Dig to Core — full dive to the core (separate soak, simulator "ORIXArcade Dig") |
5,264 |
88 |
59.6 fps |
16.7 ms |
16.7 ms |
123 ms |
59.6 |
The automatic effect scaling never engaged (final effect scale 1.0 in every log). The isolated long frames coincide with scene start-up and with screenshots being taken during the capture run.
| State |
Footprint |
Peak |
| Lobby, 8 s after launch |
38 MB |
38 MB |
| Domino City, level 15 looping, 25 s |
40 MB |
41 MB |
| Dig to Core, looping dives with a maxed drill, 25 s |
48 MB |
79 MB |
| Traffic Mayhem, Twin Junction endless, 25 s |
53 MB |
84 MB |
# Tests
| Suite |
Result |
Command |
| Unit (XCTest, hosted) |
88 executed, 0 failures, 1 skipped (opt-in tuning sweep) in 47.3 s |
Scripts/build.sh test |
| UI (XCUITest) |
6 executed, 0 failures in 216 s |
Scripts/build.sh uitest |
| Domino City reference solutions in the real app |
15 / 15 levels cleared, 3 stars each (raw/domino-autoplay.json) |
-ORIXGame dominoCity -ORIXLevel 1 -ORIXAutoplay YES |
| Traffic fairness |
bot that only launches when the predictor says "safe": 7 simulated minutes × 3 seeds × 3 junctions, and 25 minutes on Twin Junction, zero hearts lost |
part of the unit tests |
| Dig generation |
guaranteed route verified across seeds and chunks; 600 s simulated run keeps ≤ 6 chunks loaded |
part of the unit tests |
# Not measured
| Metric |
Status |
| Anything on a physical iPhone (frame rate, 120 Hz, haptics, audio output, thermals, battery) |
not measured — no device was attached |
| Cold launch time |
not measured |
| Instruments traces (Time Profiler, Allocations, Energy) |
not measured |
| Small-screen layouts (iPhone SE class), VoiceOver pass, Dynamic Type extremes |
not measured |
| Airplane mode on a device |
not measured on a device; the binary links no networking code path of its own and declares no network entitlement or permission |
| Human playtesting (feel, difficulty curve over 5–10 minute sessions) |
not done — all gameplay runs were bot-driven |