Thanks to OhGodAGirl I found this thread, now I have some ques to ask, Kristy mentioned that this FPGA will be reducing the load from GPU for GPU demanding Keccak and others part (NonFP ) which most known algorithm uses - ETH/Equi/(ProgPOW if it becomes reality). That means if this can do say Keccak part at 15W then the GPU Power usage will drop significantly - as mentioned by Kristy again. So can you confirm in details? Also your in-house test say for a OC'ed RX580 how much power can you save?

I can understand that exact values are not possible at this point of time, but can you share estimated % on say a RX580? If I buy.... the power part will be the main driver and will help me decide. Would appreciate if you can share some info on this.

The 3 models are present I can understand - what's the diff between using all these three - if you don't talk about the speed then how they differ? Only Diff between these 3 I read was on memory side, how well they will perform from each other?