spinning up GPUs
0%
Switch with a flag. Same binary, same OC controls, same dashboard — the core splits the work.
--coin pearl--coin csd--coin midstate--coin alphanumeric--coin parano1d--coin btx--coin quantusNo config files, no tuning marathon. Point it at a pool and it hashes.
A hashrate counts for nothing until someone else gets the same number on the same card.
Miners publish their OC settings. Others with the same GPU apply them, measure, and report what they actually got. Every ranking is built from those measurements — never from the submitter's own claim.
| Card | Core | Mem | W | TH/s | GH/W |
|---|---|---|---|---|---|
| 01RTX 5090 ✓ | 2400 | 7000 | 555 | 392.0 | 706.4 |
| 02RTX 5080 ✓ | 2955 | 14300 | 390 | 242.9 | 622.7 |
| 03RTX 5070 Ti ✓ | 2805 | 13800 | 280 | 193.7 | 691.8 |
| 04RTX 4080 SUPER ✓ | 2235 | 11250 | 260 | 178.0 | 684.5 |
| 05RTX 4080 ✓ | 2235 | 10800 | 250 | 168.6 | 674.2 |
| 06RTX 4070 Ti ✓ | 2625 | 10250 | 285 | 156.7 | 549.7 |
Lower output, lowest bill, coolest card.
Good output without pushing the card.
Highest hashrate the card will hold.
List several --url flags and PeakMiner fails over on its own. BTX runs on LuckyPool. QUAN runs on Kryptex, QuanPool or LuckyPool. NOID runs on InnovLab or suprnova at 3%, or solo against your own peakpool at 0%.
github.com/peakminer/peakminer/releases/download/v2.18.1/peakminer-2.18.1.tar.gz
coin = pearl · template = %WAL%.%WORKER_NAME%
per-GPU stats land in the HiveOS dashboard
Release announcements, tuning threads, and a 30-second feedback form that lands straight on our board.