A tricky pipeline of a matrix multiplier with 3 BRAMs 4 DSPs, calculating in 4 threads. - View it on GitHub