I had this kind of cool solution where one chamber was 1/2P and one chamber was 6P using halt, but of course just being 2P all-around leads to much better results. 1P would be even better of course, but this waste-handling method doesn't seem to work at 1P due to the part where the two-atom pair gets snuggled between the two checkmarks.

Now with more variable waste to save instructions and cycles.