New AI tool PPAPlace optimizes chip placement for better performance

Researchers have developed PPAPlace, a novel AI-driven system designed to optimize chip placement for improved performance, power, and area (PPA). Unlike traditional methods that focus on half-perimeter wirelength (HPWL), PPAPlace uses a dual-stream predictor that combines graph attention and spatial convolution to predict post-route timing metrics like worst negative slack (WNS) and total negative slack (TNS). This approach allows gradients from these timing metrics to be fed back into the placement process, leading to significant improvements in WNS and TNS compared to existing baselines. AI

IMPACT This research could lead to more efficient chip designs by improving the accuracy of AI-based placement tools in predicting and optimizing critical timing metrics.

RANK_REASON The cluster contains a research paper detailing a new AI method for chip placement optimization. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• ChipBench
• graph attention network
• half-perimeter wirelength
• PPAPlace
• Projected Gradient Descent
• total negative slack
• worst negative slack

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →