https://www.amd.com/en/products/cpu/amd-ryzen-7-1700x

close to my Rysen 1600, and yeah max perf is --auto -t 4 --archi ryzen
Heavy is the algo where, honnestly, i beat xmrig by the lowest factor, i'm less than 1% better. on 64 bits aes. on 32 bits and non aes i obliterate it.

edit: all code done for dualshare, now entering test phase.
perfs on one thread, with compared to without doublehash

core2 64 bits : +5%
core2 32 bits +11%
ryzen 64 bits : +92%
ryzen 32 bits +99%

The ryzen pipelining is impressive...