Same scene, camera, geometry and renderer. AVBD draws from its solver buffer; Rapier poses are copied into the same rendering layout. Reports include completed work time and animation-frame intervals. This is one integration workflow, not a universal engine benchmark. Keep the tab visible.