|Session Name||NVIDIA Aftermath: A New Way of Debugging Crashes on the GPU (Presented by NVIDIA)|
|Track / Format||Programming|
With rendering APIs increasing in complexity we need to find a better way of debugging issues that present on the GPU. Currently, there's no way to debug a GPU crash after the fact, which is not only a problem during development of a game, but also after release. Many game developers are collecting telemetry about CPU crashes - but what about GPU crashes? This session presents a solution to that problem.