Good point, and agreed the landing page is a bit sensational. I mentioned it elsewhere but between MatX and cuPy we see a 3-4x performance difference on average

Good point, and agreed the landing page is a bit sensational. I mentioned it elsewhere but between MatX and cuPy we see a 3-4x performance difference on average. The gap tends to widen with more complex workflows where compile-time kernel fusion gives more improvements compared to something like a single GEMM.