Automating Constraint-Aware Datapath Optimization using E-Graphs
Numerical hardware design requires aggressive optimization, where designers exploit branch constraints, creating optimization opportunities that are valid only on a sub-domain of input space. We developed an RTL optimization tool that automatically learns the consequences of conditional branches and exploits that knowledge to enable deep optimization. The tool deploys custom built program analysis based on abstract interpretation theory, which when combined with a data-structure known as an e-graph simplifies complex reasoning about program properties. Our tool fully-automatically discovers known floating-point architectures from the computer arithmetic literature and out-performs baseline EDA tools, generating up to 33% faster and 41% smaller circuits.
Presentation (EGRAPHS Automating Constraint-Aware Datapath.pdf) | 392KiB |
Sun 18 JunDisplayed time zone: Eastern Time (US & Canada) change
09:00 - 11:00 | |||
09:40 20mTalk | Automating Constraint-Aware Datapath Optimization using E-Graphs EGRAPHS Samuel Coward Imperial College London, UK / Intel Corporation, George A. Constantinides Imperial College London, UK, Theo Drane Intel Corporation, USA Pre-print File Attached | ||
10:00 20mTalk | egglog In Practice: Automatically Improving Floating-point Error EGRAPHS | ||
10:20 20mTalk | Optimizing Stateful Dataflow with Local Rewrites EGRAPHS Shadaj Laddad University of California at Berkeley, Conor Power University of California at Berkeley, Tyler Hou University of California, Berkeley, Alvin Cheung University of California at Berkeley, Joseph M. Hellerstein University of California, Berkeley Pre-print File Attached | ||
10:40 20mTalk | Egg-smol Python: A Pythonic Library for E-graphs EGRAPHS Link to publication Pre-print |