Dev Update 2026-01-12

Development
1

Dev Update

Last Week by the Numbers

  • neptune-core
    • Issues: 1 raised.
    • Pull Requests: 1 opened, 1 merged.
  • triton-vm
    • Issues: 1 raised.
    • Pull Requests: 1 opened, 1 merged.

Stand-up Summary

  • Thorkil Værge
    • Last Week: Merged the UTXO index to enable quick wallet rescanning; proposed shifting the implementation to reference block heights (U64) instead of hashes for disk efficiency. Created a tracking issue for necessary wallet endpoints.
    • Coming Week: Implement block height changes for the UTXO index, look into GPU prover implementations, and review the newly developed SDK.
  • Alan Szepieniec
    • Last Week: Reached 95% completion on libP2P integration. Managed significant CI challenges related to unmaintained Level DB wrappers and advocated for moving to a different database.
    • Coming Week: Review PRs for STIR and proof of payment/reserves; finalize libP2P branch and CI fixes; prepare a presentation for Starkware on transaction architecture.
  • Ferdinand Sauer
    • Last Week: Integrated STIR to replace FRI to facilitate faster future provers. Raised concerns regarding Triton VM regressions on the master branch and provided insights on the pitfalls of privacy “dummies.”
    • Coming Week: Address feedback and review cycles for the STIR PR; collaborate on GPU prover code reviews and perform maintenance on Triton PRs.
  • Dan Da
    • Last Week: Handled meeting logistics and Signal coordination.
    • Coming Week: Commencing a hiatus for personal projects; remains reachable via Signal and email for urgent matters.
  • [Redacted]
    • Last Week: Successfully integrated with new privacy coins; open-sourced a NodeJS offline server for MEXC listing requirements; implemented sub-addresses to optimize withdrawal performance and UTXO scanning.
    • Coming Week: Unify and clean up overlapped code for RPC and UTXO indexing; share GPU prover resources with the team.

Technical Discussion

The core technical focus this week was the transition from FRI to STIR within the Triton VM. While this enables significantly faster proving times in the long term, it was noted that the current master branch faces a temporary regression where it cannot yet verify its own programs recursively.

The team also debated the merits of the UTXO index design. By referencing block heights instead of 40-byte hashes, the project can save significant disk space without losing features. There was also a deep dive into GPU prover performance; current implementations show the ability to complete proofs in roughly 52 seconds at a very low operating cost ($1.50/hr), compared to the previous 80-175 second benchmarks on CPUs.

Finally, the discussion covered sub-addresses and input selection. While sub-addresses can speed up withdrawals by simplifying receiver ID queries, the team discussed whether archival node assumptions or specific mutator set indices might achieve similar performance gains for wallets.

Updates and Announcements

2 Likes
2

MEXC？

Is this work done by XNT on mexc integration, and then merged into NPT?

3

FYI This post had been flagged by the system for suspicious activity. I see only a legitimate question so I approved :slight_smile:

The party in question was named [Redacted] out of a) too little time to sync about disclosure preferences before this note was published; and b) a preference for erring on the side of privacy. I will let the party in question out themselves if they want to; otherwise, I will continue to err on the side of privacy.

I can say this – and I hope that’s enough to answer the bulk of your question:

  • To my best knowledge, the spin-off (XNT) is moving forward with the MEXC listing as they announced.
  • That listing almost certainly requires significant implementation support on the side of the full node software (xnt-core).
  • Given the similarity between the two repositories, we hope that neptune-core can benefit from the same upgrades. In fact, we are eager to support exchange integration in the abstract; and to support it urgently in order to avoid duplicate work later.
  • The dev standup meeting is not particularly private but in order to be honored with an invitation you need to express a credible desire and expectation to contribute to Neptune Cash.
1 Like