ForgeMiner is a fast, native NVIDIA GPU miner for Pearl (PRL) and KawPow (Ravencoin, Quai, Neurai), with a dedicated kernel per GPU generation and built-in overclocking

ForgeMiner A native NVIDIA GPU miner for Pearl (PearlHash) and KawPow (Ravencoin, Quai, Neurai), talking to the GPU directly through the CUDA Driver API — no Python, no WSL, no extra runtimes. Pros: Dedicated kernel per GPU generation (Pascal, Turing, Ampere, Ada, Blackwell), auto-selected at launch. Near-zero CPU load — direct CUDA Driver API, no Python/toolkit, runs well on weak hosts and crowded rigs. Built-in overclocking, power limit and fan control (fixed speed or temperature curve) — no third-party OC tool needed. Multi-pool with automatic reconnect and fail-over. HiveOS ready — drops straight into a Custom miner slot. Cons: Closed-source; releases are published only on GitHub, Telegram and Discord. NVIDIA only — no AMD, Intel, or CPU mining. List of Coins and Algorithms: Pearl (PRL) — PearlHash Ravencoin (RVN) — KawPow Quai (QUAI) — KawPow Neurai (XNA) — KawPow Xelis (XEL) — XelisHash BTX (btx.dev) — BTX Conflux (CFX) — Octopus. Needs about 8.6 GB of free video memory and that grows as the chain advances, so 8 GB cards cannot mine it and 12 GB or more is comfortable. Fees Pearl 2% KawPow (RVN / QUAI / XNA) 0.7% Xelis 1% BTX 1% Conflux 1.5% Supported GPUs NVIDIA Pascal (GTX 10, P104-100), Volta (Titan V), Turing (RTX 20), Ampere (RTX 30), Ada (RTX 40), Blackwell (RTX 50), and CMP mining cards. RTX 20-series and CMP cards need driver 545+.

  • Supports NVIDIA

  • Coins: PRL, RVN, QUAI, XNA, BTX, XEL, CFX

  • Pros: Dedicated kernel per GPU generation (Pascal, Turing, Ampere, Ada, Blackwell), auto-selected at launch.

  • Near-zero CPU load — direct CUDA Driver API, no Python/toolkit, runs well on weak hosts and crowded rigs.

Protected & verified downloads

Features

  • Supports NVIDIA
  • Coins: PRL, RVN, QUAI, XNA, BTX, XEL, CFX
  • Pros: Dedicated kernel per GPU generation (Pascal, Turing, Ampere, Ada, Blackwell), auto-selected at launch.
  • Near-zero CPU load — direct CUDA Driver API, no Python/toolkit, runs well on weak hosts and crowded rigs.
  • Built-in overclocking, power limit and fan control (fixed speed or temperature curve) — no third-party OC tool needed.
  • Multi-pool with automatic reconnect and fail-over.

System Requirements

Windows

  • Windows 10 / 11 (64-bit)
  • 4 GB RAM
  • 200 MB available space

macOS

  • Windows 10 or later
  • 4 GB RAM or more
  • 500 MB available space