Files
papers/zoo-zchain-zkp/sections/08-empirical-evaluation.tex

43 lines
1.1 KiB
TeX

\section{Empirical Evaluation}
\subsection{Throughput}
On a cluster of 5 Zoo validators (each with NVIDIA H100 80GB):
\begin{table}[h]
\centering
\begin{tabular}{lrrr}
\toprule
\textbf{Operation} & \textbf{Throughput} & \textbf{Latency} & \textbf{Gas} \\
\midrule
STARK verify (simple) & 12,000/s & 1.1s & 100K \\
STARK verify (complex) & 3,200/s & 1.8s & 500K \\
Groth16 verify & 28,000/s & 0.4s & 230K \\
PLONK verify & 18,000/s & 0.6s & 350K \\
Receipt generation & 8,500/s & 1.4s & 50K \\
PQZ export (STARK$\to$Groth16) & 450/s & 8.2s & 800K \\
\bottomrule
\end{tabular}
\caption{Z-Chain verification throughput on GPU validators.}
\end{table}
\subsection{Comparison}
\begin{table}[h]
\centering
\begin{tabular}{lcccc}
\toprule
& \textbf{Z-Chain} & \textbf{=nil;} & \textbf{Polygon} & \textbf{StarkNet} \\
\midrule
PQ-safe & Yes & No & No & Partial \\
Multi-system & 3+ & 1 & 1 & 1 \\
Receipts & Native & No & No & No \\
FHE hybrid & Yes & No & No & No \\
GPU accel. & Yes & No & No & No \\
\bottomrule
\end{tabular}
\caption{Z-Chain vs existing ZK infrastructure.}
\end{table}