mirror of
https://github.com/lux-apps/papers.git
synced 2026-10-30 18:15:21 +00:00
zeekayandHanzo Dev
178febad68
Every rendered "T-Chain" / "ThresholdVM" / "teleportvm" reference across the paper corpus is rewritten to the correct canonical chain by concern: threshold signing / MPC / bridge custody / key recovery -> M-Chain (MVM); FHE / encrypted compute / threshold decryption -> F-Chain (FVM); teleport / cross-chain messaging -> B-Chain (bridgevm); consensus Pulsar stays Q-Chain. Section lux-trust-kernel/03-t-chain-threshold-governance.tex renamed -> 03-f-chain-encrypted-policy.tex (with \input + xref fixes). Braided "T-Chain" sections split by concern (M vs F). Only lp-020's chain-separation rationale and the shared lux/chains/thresholdvm library path retain the name, to record the removal. All changed papers rebuilt (pdflatex, rc=0, no undefined refs). ยง09 of lux-dao-governance-framework reframed: fabricated "production deployment" metrics ($50M/$240k/$12M, survey %s, "as of Oct 2024") relabeled as illustrative design configurations with no invented replacements; reality noted (DAO treasury is a 1/1 Safe); conclusion echo + one "first production" overclaim softened. This commit also captures concurrent corpus workstreams on the same tree. Co-authored-by: Hanzo Dev <[email protected]>
17 lines
2.1 KiB
TeX
17 lines
2.1 KiB
TeX
\section{Conclusion}
|
|
\label{sec:conclusion}
|
|
|
|
LP-137 specifies the GPU mode shared by the cryptographic primitives in the Lux 9-chain topology. The substrate is a byte-equality \emph{contract} across three backends (Metal, CUDA, WGSL): a GPU kernel is admitted to the production dispatch path only once it is byte-equal to the same Go reference on $N=1000$ random inputs, and a kernel that fails byte-equality, lacks a validated backend KAT, or is deliberately left unwired (e.g.\ a KAT-clean lattice-NTT kernel whose single-shot GPU path is slower than the CPU body) falls back to the CPU canonical. PQ signature \emph{verify} at the P3Q slot runs on the pure-Go/CPU path in production; there is no production GPU kernel for PQ verify. For every kernel that is on a production hot path, the substrate satisfies the GPU-residency invariant: no chain-local hot path leaves attested GPU memory.
|
|
|
|
The per-algorithm formal specs live in their own LPs. LP-137 (this paper) carries no per-algorithm content beyond a pointer index (\S\ref{sec:per-algorithm-spec}). This separation lets the substrate evolve (new GPU backend, new attestation profile, new dispatcher heuristic) without re-opening per-algorithm specs, and lets each algorithm evolve (new KAT vectors, new constant-time refinement) without re-opening the substrate.
|
|
|
|
\paragraph{Future work.} Three substrate-level extensions are tracked:
|
|
|
|
\begin{enumerate}
|
|
\item Multi-GPU dispatcher: extend the determinism harness to verify byte-equality across heterogeneous GPU backends in the same cluster.
|
|
\item AMD ROCm and Intel oneAPI backends: the same byte-equality contract applied to two more shader languages, joining Metal/CUDA/WGSL.
|
|
\item Hardware-attested GPU memory pinning: integrate with the LP-127 attestation envelope so that the determinism harness can require an attestation receipt for every shipped GPU build.
|
|
\end{enumerate}
|
|
|
|
The 1-LP-per-algorithm decomposition (LP-026, LP-029, LP-066, LP-069, LP-070--077, LP-110--126, LP-131, LP-146--159) is the canonical entry point for any consumer who needs a specific primitive's formal spec. LP-137 is the substrate that holds them together.
|