I tried some, when I was checking what would work if I really had to scrounge bullets someday. The only .22 mould I had was a Lyman 225415, which has been made in 50 and 55 grain forms over the years. Mine is a 55.
Casting is full of things that "everybody knows", and one is that casting for any .22 centerfire is the supreme test of patience. Maybe I was lucky, but I got it to work with less trouble than people made out.
I wanted to keep it 1800-2000 fps. The trick is keeping velocity (and pressure) that low while getting enough gas pressure at the port to run the action. I had read that AA 2015BR was the ticket, but the two loads I tried ran faster than I wanted (2300-ish) and didn't function. Of course I can't find the notes now, and forget what powder it was, but the next powder worked. It ran 2030 fps and cycled fine.
I only shot at 50 yards because I was just trying to get function at that point. It shot around 2" there, but was right on the same POI as M193 at that range.
One thing I've found when testing cast loads in semiauto rifles is mag capacity will influence function. A load that works with five rounds in the magazine may not work with 10 or 20 due to increased bolt drag. I did not test these with more than five rounds because I didn't load many.
I'll dig up my data later if interested.