NetSmith: An Optimization Framework for Machine-Discovered Network Topologies
Conor Green, Mithuna Thottethodi
TL;DR
NetSmith addresses the challenge of designing high-performance network topologies for general-purpose, shared-memory multicore systems, where expert-designed interposer networks have dominated. It introduces a MILP-based topology generation framework that optimizes either latency (average hops) or sparsest-cut throughput under realistic hardware constraints, producing irregular but deadlock-free NoI topologies. Across 4×5, 20-router interposer configurations and larger scales, NetSmith consistently outperforms prior expert designs and previous optimization-based NoC methods in both analytical metrics and system simulations, achieving up to 11% application speedup on PARSEC workloads. The results demonstrate that automatic topology discovery can yield substantive performance gains with feasible compute times, while maintaining compatibility with established routing and VC-allocation strategies, suggesting practical adoption for interposer-based, chiplet-structured systems.
Abstract
Over the past few decades, network topology design for general purpose, shared memory multicores has been primarily driven by human experts who use their insights to arrive at network designs that balance the competing goals of performance requirements (e.g., latency, bandwidth) and cost constraints (e.g., router radix, router counts). On the other hand, there have been automatic NoC synthesis methods for SoCs to optimize for application-specific communication and objectives such as resource usage or power. Unfortunately, these techniques do not lend themselves to the general-purpose context, where directly applying these previous NoC synthesis techniques in the general-purpose context yields poor results, even worse than expert-designed networks. We design and develop an automatic network design methodology - NetSmith - to design networks for general-purpose, shared memory multicores that comprehensively outperform expert-designed networks. We employ NetSmith in the context of interposer networks for chiplet-based systems where there has been significant recent work on network topology design (e.g., Kite, Butter Donut, Double Butterfly). NetSmith generated topologies are capable of achieving significantly higher throughput (50% to 75% higher) while also reducing average hop count by 8% to 13.5%) than previous expert-designed and synthesized networks. Full system simulations using PARSEC benchmarks demonstrate that the improved network performance translates to improved application performance with up to 11% mean speedup over previous NoI topologies.