Kudos to the weeklies team on a well-designed sequel to Sophick Mercury!
Hoping to hit at least one global pareto this year.

Unfortunately 17C seemed too much of a head-scratcher to bother with, so I settled for 18C on the fast end.
This uses the rate 2 core loop from a 19C solution, and the first fire atom used as a catalyst is hastily channeled through  additional glyphs tacked on to the top. But using this method, if the core loop is on lowest latency, there will be not enough cycles to complete and output the sixth product for 17C.
Actual min-cycle solutions probably have to use previous outputs as a duplicator catalyst instead. Looking forward to see solutions from those who attempted this.