TimingLLM: A Two-Stage Retrieval-Augmented Framework for Pre-Synthesis Timing Prediction from Verilog

TL;DR

TimingLLM is a novel two-stage retrieval-augmented LLM framework that predicts post-synthesis timing metrics directly from Verilog, enabling faster and adaptable timing estimation for RTL design iterations.

Contribution

It introduces a new two-stage LLM pipeline with a lightweight structural-timing cue system and a learned steering vector, improving speed and adaptability over prior methods.

Findings

Achieves R_WNS=0.91 and R_TNS=0.97 on VerilogEval with low MAPE.

Runs 1.3-1.6 times faster than previous approaches.

Can be adapted to new technology libraries with minimal retraining.

Abstract

Early, tool-free prediction of post-synthesis timing remains a key obstacle to rapid RTL iteration. We introduce TimingLLM, a two-stage retrieval-augmented LLM pipeline that estimates worst negative slack (WNS) and total negative slack (TNS) directly from Verilog. Stage 1 is a fine-tuned LLM that acts as a compact post-synthesis timing oracle, producing path-level arrivals/required times that are summarized into lightweight structural-timing cues (e.g., bag-of-gates counts, critical-path depth, gate-type patterns). Stage 2 is an LLM-based regressor that predicts WNS/TNS and applies a learned diagonal steering vector at the last transformer block, computed from the k nearest timing-labeled modules in a disjoint retrieval bank. On VerilogEval, TimingLLM attains R_WNS = 0.91 (MAPE 12%) and R_TNS=0.97 (MAPE 16%) while running 1.3-1.6 times faster than prior methods. Training uses a new 60k-module Verilog corpus with synthesis reports, which we will release. After training once, TimingLLM can be adapted to new technology libraries and PVT corners by refitting only a small regression head on 1000 labeled modules per setting, consistently outperforming state-of-the-art baselines.