[Libre-soc-dev] daily kan-ban update 28apr2022

[Jacob Lifshay]

Well, I got the HDL for goldschmidt division to work -- it probably
needs to be a FSM since the full combinatorial 128/64->64 div/mod
circuit gives 700k cells from yosys synth:
(a part of that is probably that yosys tried to convert the 3kbit rom
to gates rather than leaving it as a rom)

   Number of cells:             697167
     $_ANDNOT_                  199345
     $_AND_                      11668
     $_MUX_                       2478
     $_NAND_                      1330
     $_NOR_                     127038
     $_NOT_                      29598
     $_ORNOT_                    19964
     $_OR_                       84390
     $_XNOR_                     73265
     $_XOR_                     148091

This is with the parameters selected optimizing for latency at the
expense of multiplier size.

Jacob