Bitcoin Forum
August 18, 2026, 10:33:46 AM *
News: Latest Bitcoin Core release: 31.1 [Torrent]
 
   Home   Help Search Login Register More  
Pages: [1]
  Print  
Author Topic: Independent PoW review requested: can Soveroot v1 run with half memory?  (Read 12 times)
Emily.darryl (OP)
Newbie
*
Offline

Activity: 6
Merit: 0


View Profile
August 16, 2026, 03:31:02 PM
 #1

Soveroot PoW v1 — independent technical review requested

Soveroot is an experimental Bitcoin-derived protocol project seeking independent review of its isolated, non-consensus PoW v1 candidate.

The main challenge asks whether the exact candidate output can be computed using no more than 131,072 bytes of total mutable attack state—half the standard 262,144-byte scratchpad.

Two evaluation tracks
  • Screening: eight fresh cases, a 5,000,000-operation ceiling, with partial exact progress retained.
  • Completion: eight fresh cases, no artificial operation ceiling, with exact canonical proof required.

The repository includes
  • A byte-level candidate specification.
  • Python and independent C++ implementations.
  • Public qualification cases and test vectors.
  • Machine-readable challenge, submission, and results formats.
  • Commit-reveal derivation of fresh cases after source freeze.
  • Complete memory and operation-accounting requirements.
  • An evaluator runbook for network-disabled execution and physical-memory review.

Important scope and safety notes

This is not a live network, token sale, investment offer, bounty, or security claim. The candidate is not in consensus and is not safe for monetary use. Failed, partial, invalid, ineligible, and successful results will all be preserved. A lack of successful attacks would not prove security.

Research call and discussion
GitHub issue #49: independent researchers wanted

Repository
Soveroot decentralized-currency-protocol

Independent attackers, mining-software developers, compiler/performance reviewers, FPGA/ASIC analysts, and memory-accounting reviewers are welcome. Pseudonymous participation is accepted when the code, build instructions, accounting, and results are public and reproducible.

Technical criticism is welcome, especially around attack-state accounting, time-memory tradeoffs, implementation ambiguity, and whether the proposed challenge can produce meaningful evidence.
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!