Competitor benchmark for AppDimens Dynamic KMP: measures Dynamic 1.0.1 vs SDPS 3.1.6 vs Lib #2 on every supported platform with identical methodology (same composition, identical warm-up, 9 samples × 50,000 iterations, per-sample order rotation, anti-DCE checksums, chunked at 5,000 ops/frame).
| Target | Entry point |
|---|---|
| Android | benchlab-android module (release APK + R8) |
| JVM desktop | jvmMain/kotlin/com/example/benchlab/main.kt |
| iOS | iosMain/kotlin/com/example/benchlab/MainViewController.kt |
| macOS | macosMain/kotlin/com/example/benchlab/main.kt |
| Web / wasmJs | wasmJsMain/kotlin/com/example/benchlab/main.kt |
Lib #2 is isolated behind an expect/actual bridge (Lib2Bridge) so the macOS target — which Lib #2 does not publish — still runs the full comparison with an equivalent formula (min(w,h)/300, byte-for-byte the same rule on every platform).
# Android (device)
./gradlew :benchlab-android:installRelease
adb shell am start -n com.example.benchlab/.MainActivity --ez AUTO_START true
adb logcat -s BENCHLAB
# Desktop (JVM)
./gradlew :benchlab:run --args="--auto-start"
# Web
./gradlew :benchlab:wasmJsBrowserDistribution
# serve build/dist/wasmJs/productionExecutable, open in a browser, results print to console
# Native frameworks
./gradlew :benchlab:linkDebugFrameworkIosArm64 :benchlab:linkDebugExecutableMacosArm64Results (probe medians, engine medians, legacy T1/T2/T3, device info) are printed to the console / logcat (BENCHLAB tag). Latest numbers: PERFORMANCE.md · PERFORMANCE-COMPARATIVE.md.