<whitequark[cis]> "you could use xoroshiro for..." <- This algorithm sounds very interesting— I'm finding discussion *about* xoroshiro, and I'm finding code, but I'm having trouble finding like, a clear description or a block diagram.
I think I could convert this verilog code to amaranth line by line:
https://github.com/ZiCog/xoroshiro/blob/master/Xoroshiro32PlusPlus.v
But I definitely would not have a clear sense of what I was doing, lol
With an LFSR it seems you need to do multiple rounds to get a particularly random bitmask… how many rounds of xoroshiro do you need to get "random looking" values?
(Or is that why you calculate a "jump function"?)