ZK-Tracer: A High-Performance Heterogeneous Accelerator for Zero-Knowledge VM Trace Generation

📅 2026-05-25
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
This work addresses the critical performance bottleneck in zero-knowledge proof systems caused by front-end trace generation in zero-knowledge virtual machines (zkVMs). It presents the first heterogeneous hardware acceleration architecture tailored specifically for zkVM front-end execution, featuring coordinated processing units for main and parallel permutation traces alongside a lightweight instruction set extension to enable fine-grained hardware-software co-design. Unlike prior efforts that focus exclusively on back-end proving, this architecture targets the underexplored front-end phase. ASIC implementation results demonstrate a 1,829× speedup in trace generation over a high-performance multicore CPU; when integrated with existing back-end accelerators, the end-to-end proof generation achieves a 963× performance improvement.
📝 Abstract
Zero-knowledge virtual machines (zkVMs) are a key technology for driving the large-scale adoption of zero-knowledge proofs (ZKP), but their performance bottlenecks severely limit their practicality. While current hardware acceleration research has exclusively focused on backend proving, we identify that the frontend execution and trace generation phase is rapidly emerging as the new system bottleneck. To address this challenge, we propose ZK-Tracer, the first hardware accelerator architecture specifically designed for the zkVM frontend. ZK-Tracer features a novel heterogeneous design comprising a Main Trace Unit and parallel Permutation Trace Units. It exposes a fine-grained interface to the host software through a lightweight instruction set extension, enabling efficient task offloading. Our ASIC implementation results demonstrate that ZK-Tracer achieves up to 1829x speedup in trace generation over a high-performance multi-core CPU. When integrated with existing backend proving accelerators, it delivers a remarkable 963x end-to-end performance improvement for the entire ZKP system.
Problem

Research questions and friction points this paper is trying to address.

zero-knowledge virtual machines
performance bottleneck
trace generation
frontend execution
zero-knowledge proofs
Innovation

Methods, ideas, or system contributions that make the work stand out.

zkVM
hardware accelerator
trace generation
heterogeneous architecture
zero-knowledge proof
🔎 Similar Papers
No similar papers found.
J
Jieran Cui
School of Integrated Circuits, Southeast University, China; National Center of Technology Innovation for EDA, China
Z
Zhengkai Wen
School of Integrated Circuits, Southeast University, China; National Center of Technology Innovation for EDA, China
Haowen Fang
Haowen Fang
Synopsys
Machine LearningNeuromorphic ComputingElectronic Design AutomationFPGA
Y
Yinan Zhu
School of Integrated Circuits, Southeast University, China; National Center of Technology Innovation for EDA, China
J
Jia Xiong
School of Integrated Circuits, Southeast University, China; National Center of Technology Innovation for EDA, China
C
Cheng Ni
School of Integrated Circuits, Southeast University, China; National Center of Technology Innovation for EDA, China
M
Mingchi Zhang
School of Integrated Circuits, Southeast University, China; National Center of Technology Innovation for EDA, China
Nan Guan
Nan Guan
City University of Hong Kong
Cyber-Physical systemsEmbedded systemsReal-time systems
Xi Wang
Xi Wang
Southeast University
Computer ArchitectureMemory systemCompilerEDA