31 bits, DP 6
[1000] Ops: avg 48106 = 1.036 * sqrt(b) min 1164 max 119897 dp_ovh: 96.0 mul: 8.0
Stored footprints: avg 1719 min 44 max 4389
39 bits, DP 8
[20] Ops: avg 698950 = 0.942 * sqrt(b) min 219204 max 1848433 dp_ovh: 384.0 mul: 9.35
Stored footprints: avg 6410 min 1987 max 16978
47 bits, DP 10
[2] Ops: avg 11267030 = 0.950 * sqrt(b) min 7221075 max 15312985 dp_ovh: 1536.0 mul: 10.50
Stored footprints: avg 25820 min 16542 max 35099
No cycles. No "hey, let's create kangaroos and pretend they don't count as a group operation".
This only works for secp256k1. It is basically 100% faster than the usual approach.
I hope no one actually believes they can compete with RetiredCoder aka 3Emi. It looks like a billionaire's hobby, not a contest. It costs at least 500K to break 135, but he doesn't do it for the money, and I won't be a sheep.