@User re your questions above, see #silicon-validation. I've been testing an early OpenRAM test chip predating MPW1 containing a 32 x 256 OpenRAM instance with very slow top level routing between it and the I/O pads