I started out by tring to make a latency solution, I'm pretty sure 12 is theoretical minimum, but for a design that can work with 2 pipelines, I couldn't get it lower that 14.

I tried cutting the rate per pipeline down to 4, but arm 8 and 9 (24 and 25 in the bottom pipeline) are a problem for that, I also seriously doubt if it's possible without increasing the per pipeline latency to 16 or more (which is needed to decrease the total cycle count).

I hope I did well, I had the base of this design pretty quick, so that might be a bad sign.

I sadly won't be able to catch the stream cuz I have dnd, so pls let me know how I did :3