You can define environments and instead of having to do the equivalent of cinematic set design micromanagement, let the engine come up with how the room looks and sounds and 'feels'.Īnd with respect to physics, if you model the structure of the game world, you can let the player use different forces to literally tear it apart. You can then allow modifications and let the engine code on the GPU handle the changes.
wav files for different firearms, for example, you can define a heavy machinegun or small pistol loosely and then fine-tune the output a bit, but do all of that on the GPU. The more data you can put into the models, the more accurate it is instead of loading. everything has physical properties that affect structure, reflectivity, and acoustics. example: Ray tracing can be considered to be a form of 'physics' calculation, as can acoustic wave bounces as game environments get more sophisticated especially with the advent of VR, putting all of the 'simulation' data on the GPU might be adventageous. With respect to physics, running it on a GPU when trying to do more than basic stuff can make sense and probably will make sense at some point. To a degree- there's still plenty of usefulness for hardware offload though, of course.
0 kommentar(er)
