I spot a
windmill! 1 ...Bxg3+ 2 Kg1 Bh2+ 3 Kh1 Bxf4+ 4 Kg1 Bh2+ 5 Kh1 Bxe5+ 6 Kg1 Bh2+ 7 Kh1 Bxd6+ 8 Kg1 Bh2+ 9 Kh1 Bxc7+ 10 Kg1 Bh2+ 11 Kh1 Bxb8+ 12 Kg1
after which Black changes tack:
12 ... Bxa7+ 13 Nc5 (best) Bxc5+ 14 Rd4 (best) Bxd4 15 Be3 Bxe3 16 Rf2+ Bxf2+ 17 Qxf2#
Even after all the pawns on the g3-c7 diagonal are gone, White still guards every square on the b6-f2 diagonal, so Black has no short cut: Black's bishop must operate the windmill all the way out to b8 so that it can get to the safe square a7.
What a slaughterhouse indeed! Well done!
As requested: an explanation of why I chose the moves I did. With such a terrifyingly long selfmate I thought that for the setter to have composed it at all, the solution must surely be a simple one, with Black always making checks which White can parry only in one way. I saw that most of them put a Black piece en prise -- not a good sign.
1 Bxg3+
looked better. Though it seems to be en prise, it isn't really, because White is in check from the other piece, and capturing the discoverer is not enough to get out of discovered check. From there it was no great leap of imagination to spot the idea that drives Black's play in the first part of the solution. The long diagonal of pawns with the bishop at one end was another clue.
I must admit that I underestimated the scale of the slaughter. I wondered if Black, after capturing the f4 pawn, could check with the queen so as to force White's queen or rook down for the mate. But that didn't work.