⚡ v0.2.0 — faster GPU kernel
Reworked MatMul GPU pipeline (fused projection) that mines the V3 PoW more
efficiently — more shares per watt on the very same card. Byte-exact with network
consensus (independently validated, 0 divergence over 1000+ nonces).
Update and restart your miner — no config change needed.
- Multi-GPU fatbin: native cubins for RTX 20/30/40/50 series (sm_75 → sm_120, native Blackwell).
- Portable build (glibc 2.30, static libstdc++) — runs on Ubuntu 20.04/22.04/24.04 + HiveOS.
- 1% transparent dev fee.