Backend
PRISM: Preference-based Reward and Uncertainty-Integrated Synthesis Minimization for RTL PPA Optimization
arXiv cs.SE0 views
arXiv:2611.11051v2 Announce Type: replace-cross Abstract: This paper introduces a budget-aware, multi-objective framework (called PRISM) for RTL optimization that improves power, performance, and area (PPA) while sharply limiting expensive commercial synthesis tool calls. Starting from a synthesizable baseline and functional testbench, the method iterates through LLM-driven rewrite generation, fast compile/simulation equivalence filtering, low-cost proxy feature extraction, and selective commercial synthesis tool (Synopsys Design Compiler (DC)) evaluation. To prioritize which candidates deserve high-quality synthesis, we train a preference-based reward model using pairwise ranking supervision derived from commercial synthesis tool outcomes, and combine ensemble uncertainty with a diversity-aware acquisition rule to avoid early bias and redundant evaluations. A key advantage is that PRISM maintains a best-tradeoff set G of non-dominated designs throughout the run: G preserves the best solutions found so far, is fed back into generation and refinement, and enables consistent, cumulative improvements rather than repeatedly restarting from the baseline. Across RTLLM v2.0 modules, the approach consistently yields post-synthesis gains under fixed commercial synthesis tool call budgets, and ablations confirm that reward-guided selection and candidate dropping are central to sample-efficient progress. PRISM efficacy is evaluated across multiple technology nodes (7 nm, 15 nm, and 45 nm) using identical pipelines without retraining.
Source:https://arxiv.org/abs/2611.11051PreviousCross-Benchmark Transfer from RL on Agentic Coding TasksNextWordPress Bozuk Veritabanı Tablolarını phpMyAdmin ile Onarma
Related posts
1,016 TOPS per car: Pony.ai's Plan B for life after Nvidia
0dev.to Discuss
Your Integration Code Is Already a Workflow
0dev.to Discuss

Your Coding Agent Shouldn't Be Able to Close Its Own Ticket
0dev.to Discuss

My Sandbox Could Only Reach 3 Package Registries for 16 Weeks. This Week It Reached 8 — and One of Them Leaks Whether 2FA Is On
0dev.to Discuss
5 Python Scripts Every Tech Enthusiast Should Know for Automating Their Daily Life
0dev.to Productivity
Tencent's Hy4 Flash: A 224B Model That Costs 1/20th of Its Flagship — Here's the Trade-Off
0dev.to Productivity