mirror of
https://github.com/lux-apps/papers.git
synced 2026-11-03 05:07:46 +00:00
Retract the fabricated '581M/597ns / 100M+/s / 434M GPU / planet-scale' headline (a time.Sleep sim + x40 fudge that measured nothing) across the corpus. Flagship lux-lightspeed-dex rewritten (decomplected d-chain engine / dexvm atomic proxy / Go wrapper / 0x9010 CLOB facade; measured GB10 104-126M & M4 Max ~250M fills/s byte-parity-gated; Knuth-D division lever; Block-STM settlement; P3Q->starkfri rollup to Z-Chain). Technical papers (precompile-engineering, gpu-crypto-stack, evmgpu-benchmark, p3q-corona, smart-order-routing, fhe-mpc-hybrid) carry the measured impl + cross-ref the flagship. >1B/s framed honestly as N-node aggregation, never single-GPU. All compile.
106 lines
4.3 KiB
TeX
106 lines
4.3 KiB
TeX
\section{Architecture: 2022 Design vs.\ 2026 Implementation}
|
|
|
|
|
|
\subsection{Matching Engine}
|
|
|
|
\textbf{2022}: C++ matching engine compiled to Solidity WASM,
|
|
running on a Lux L1 chain. Order book stored on-chain.
|
|
|
|
\textbf{2026}: Multi-engine architecture
|
|
(\texttt{luxfi/dex}): Go (default), C++ (high-freq), Rust
|
|
(safety-critical), MLX (Apple Silicon). CLOB matching with
|
|
AMM fallback. Clearinghouse with margin and liquidation.
|
|
Cross-chain gateway bridging 8 chains. The GPU CLOB kernel
|
|
sustains 104--126M byte-exact \texttt{uint256} fills/s on an NVIDIA
|
|
GB10 (6.9\,ns/match) and roughly 250M fills/s on an Apple M4 Max
|
|
(4.15\,ns/match); the throughput lever is a Knuth Algorithm-D
|
|
division that replaces a 512-iteration bit-serial \texttt{uint256}
|
|
VWAP ($\sim$8$\times$), with output byte-identical to the CPU
|
|
oracle (LP-203).
|
|
|
|
\subsection{Social Trading and Copy Swaps}
|
|
|
|
\textbf{2022}: Permission-mapped liquidity aggregator smart
|
|
contracts allowing expert traders to manage investor capital
|
|
on-chain while investors retain key custody. People-Based
|
|
Portfolios (PBPs) with transparent performance stats.
|
|
|
|
\textbf{2026}: The Lux ATS (\texttt{luxfi/ats}) implements the
|
|
regulated version: SecurityToken deployments via \texttt{luxfi/standard}
|
|
on the Lux EVM, portfolio management via the exchange frontend
|
|
(\texttt{luxfi/exchange}), and social trading features
|
|
(leaderboards, copy-trade, rebalancing). The permission-mapping
|
|
concept was absorbed into the MPC wallet architecture: users
|
|
sign with their WebAuthn key share, the ATS co-signs after
|
|
compliance checks, and capital never leaves the user's
|
|
on-chain wallet.
|
|
|
|
\subsection{Decentralized Liquidity Pool}
|
|
|
|
\textbf{2022}: Uniswap V3--style concentrated liquidity with
|
|
up to 4000x capital efficiency, extended to support both
|
|
fungible and non-fungible assets in a single pool.
|
|
|
|
\textbf{2026}: The Lux DEX implements a native orderbook
|
|
(CLOB) rather than an AMM-only model. AMM pools exist as
|
|
fallback liquidity via \texttt{luxfi/standard/contracts/amm/}
|
|
(AMM V1--V4, Oracle-Mirrored AMM). The NFT liquidity pool
|
|
concept was not pursued --- the market moved toward
|
|
orderbook-based NFT trading (Blur/Tensor model) rather than
|
|
AMM curves for non-fungibles.
|
|
|
|
\subsection{Atomic Swaps and Bridge}
|
|
|
|
\textbf{2022}: HTLC-based atomic swaps with BTC, subsidized
|
|
by gateway deposits. Cross-chain interoperability via oracles
|
|
(Chainlink).
|
|
|
|
\textbf{2026}: The Z-Chain (zero-knowledge bridge) uses
|
|
homomorphic encryption for shielded cross-chain transfers.
|
|
The Lux DEX cross-chain gateway supports 8 chains natively.
|
|
\texttt{luxfi/standard/contracts/bridge/} implements the
|
|
Warp Messenger protocol for L1$\leftrightarrow$L2 asset
|
|
movement. Chainlink oracles replaced with Lux Oracle
|
|
(native, threshold-signed price feeds).
|
|
|
|
\subsection{Governance and DAO}
|
|
|
|
\textbf{2022}: LUX token governance with quadratic voting,
|
|
holographic consensus, proof-of-humanity biometric
|
|
authentication, proxy voting, and Ricardian contracts.
|
|
|
|
\textbf{2026}: DAO governance contracts exist in
|
|
\texttt{luxfi/standard/contracts/dao/} and
|
|
\texttt{contracts/governance/}. Voting is on-chain with
|
|
LUX staking. Quadratic voting and proof-of-humanity were
|
|
not implemented --- standard token-weighted voting with
|
|
delegation is the current model. Ricardian contracts were
|
|
superseded by on-chain compliance gates
|
|
(\texttt{IComplianceGate} interface).
|
|
|
|
\subsection{Security}
|
|
|
|
\textbf{2022}: SHA-3, PKI, time-delayed multisig, benign
|
|
fill, hardware wallets for all signing, homomorphic
|
|
zero-knowledge file systems.
|
|
|
|
\textbf{2026}: All realized plus post-quantum upgrades:
|
|
\begin{itemize}[nosep]
|
|
\item SHA-3 $\rightarrow$ used throughout (Keccak-256 for
|
|
EVM compatibility, SHA3-256 for non-EVM).
|
|
\item Multisig $\rightarrow$ 2-of-3 MPC
|
|
(CGGMP21 + FROST), strictly superior to multisig
|
|
(single on-chain address, no $n$-of-$m$ script
|
|
complexity).
|
|
\item Hardware wallets $\rightarrow$ HSM co-signing
|
|
(AWS KMS, GCP, Zymbit) for settlement intents.
|
|
\item Zero-knowledge file system $\rightarrow$
|
|
\texttt{PrivateStore.sol} + \texttt{/v1/base/private}
|
|
(E2E encrypted blind blob store, shipped Q1 2026).
|
|
\item FHE for front-running prevention $\rightarrow$ FHE
|
|
CRDT merge (shipped Q1 2026), encrypted orderbook
|
|
planned for Lux DEX V2.
|
|
\end{itemize}
|
|
|
|
%% ============================================================
|