OpenACMv2: An Accuracy-Constrained Co-Optimization Framework for Approximate DCiM

TL;DR

OpenACMv2 is a framework that co-optimizes architecture and transistor-level design for approximate DCiM accelerators, achieving better power, performance, and area tradeoffs while respecting accuracy constraints.

Contribution

It introduces ACCO and OpenACMv2, a novel open framework that combines architecture search and circuit sizing for approximate DCiM, enabling rapid exploration and improved PPA-accuracy tradeoffs.

Findings

Significant PPA improvements under accuracy constraints.

Fast GNN-based surrogate for PPA and error estimation.

Decoupled architecture and circuit optimization workflow.

Abstract

Digital Compute-in-Memory (DCiM) accelerates neural networks by reducing data movement. Approximate DCiM can further improve power-performance-area (PPA), but demands accuracy-constrained co-optimization across coupled architecture and transistor-level choices. Building on OpenYield, we introduce Accuracy-Constrained Co-Optimization (ACCO) and present OpenACMv2, an open framework that operationalizes ACCO via two-level optimization: (1) accuracy-constrained architecture search of compressor combinations and SRAM macro parameters, driven by a fast GNN-based surrogate for PPA and error; and (2) variation- and PVT-aware transistor sizing for standard cells and SRAM bitcells using Monte Carlo. By decoupling ACCO into architecture-level exploration and circuit-level sizing, OpenACMv2 integrates classic single- and multi-objective optimizers to deliver strong PPA-accuracy tradeoffs and robust convergence. The workflow is compatible with FreePDK45 and OpenROAD, supporting reproducible evaluation and easy adoption. Experiments demonstrate significant PPA improvements under controlled accuracy budgets, enabling rapid "what-if" exploration for approximate DCiM. The framework is available on https://github.com/ShenShan123/OpenACM.