Distributed ranges
Clients claim non-overlapping search ranges from the central service.
✓ FREE & OPEN SOURCEBITCOIN · DISTRIBUTED COMPUTING
OpenPuzzle
Distributed Bitcoin Puzzle search coordination for CPU and GPU workers.
Active projectOpenPuzzle is under active development. Public packaging and documentation continue to evolve.Project website
OpenPuzzle coordinates distributed Bitcoin Puzzle range searches. Clients claim work from a central service, run supported search engines locally and report progress back without duplicating assigned ranges.
CPU & GPU
Designed to support CPU and GPU workers with compatible search engines.
Engine integration
Works with GPU-oriented BitCrack workflows and CPU-oriented KeyHunt workflows.
Progress tracking
Workers report progress, speed and completion to the central service.
Fair scheduling
Work allocation is designed so all clients have an equal right to receive work.
Open development
The project is developed publicly with community participation in mind.
OpenPuzzle
OpenPuzzle is under active development. Public packaging and documentation continue to evolve.
What problem OpenPuzzle solves
Independent compute workers can easily duplicate the same key ranges. OpenPuzzle introduces a central coordinator that leases non-overlapping work, records progress and returns abandoned assignments to the available pool. The result is measurable shared coverage rather than isolated benchmark runs.
Designed for heterogeneous workers
The project can coordinate CPU and GPU clients using different engines and backends. Each worker reports measured progress instead of relying on theoretical hardware specifications. This makes mixed fleets useful while keeping assignment state centralized.
Public-puzzle scope
OpenPuzzle is developed for intentionally published cryptographic puzzles and other authorized research tasks. The project is not presented as a system for targeting arbitrary wallets or third-party assets.
Reliability is part of the scheduler
Lease expiry, process recovery, stop control and completion reporting are treated as core distributed-systems problems. A fast worker is only useful if the coordinator can trust which range it is running and whether that work was actually completed.