nprofile1qy2hwumn8ghj7un9d3shjtnddaehgu3wwp6kyqpqqv5atqz9k9c54q8c28kra6sfata0wk7w7x5gkrnde8vmxe5gt00q3n8x2j (nprofile…8x2j) The Galois multiplier is a fairly high priority, I need a 32 bit one for something else, will probably do 128 and others while I'm at it.
I doubt I'll be able to do single cycle 128 bit multiplies at 500 MHz on US+ so 64 Gb/s throughput on a single stream would be tricky. Being carryless helps but it's still a multiply with a bunch of partial products. Some math workarounds possible but I'd rather not go there. 10 Gb/s on US+ shouldn't need any of that.