my journey started uh largely instrumenting code to run on a single GPU. This was state-of-the-art uh when I started and that that was sort of great. We had some challenges but we had this this disc media called SATA and it sort of kept up with what we wanted to do but it didn't take long when we started instrumenting as science instrumenting code for multiGPU particularly as science began to increase the size of data we could do more.So they were doing more. So we were growing the platform capability by ad ...