Saved in:
Bibliographic Details
Main Authors: Wang, Runzhi, Sengupta, Prianka, Roman-Vicharra, Cristhian, Chen, Yiran, Hu, Jiang
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.11662
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913803868831744
author Wang, Runzhi
Sengupta, Prianka
Roman-Vicharra, Cristhian
Chen, Yiran
Hu, Jiang
author_facet Wang, Runzhi
Sengupta, Prianka
Roman-Vicharra, Cristhian
Chen, Yiran
Hu, Jiang
contents In chip design planning, obtaining reliable performance and power forecasts for various design options is of critical importance. Traditionally, this involves using system-level models, which often lack accuracy, or trial synthesis, which is both labor-intensive and time-consuming. We introduce a new methodology, called Lorecast, which accepts English prompts as input to rapidly generate layout-aware performance and power estimates. This approach bypasses the need for HDL code development and synthesis, making it both fast and user-friendly. Experimental results demonstrate that Lorecast achieves accuracy within a few percent of error compared to post-layout analysis, while significantly reducing turnaround time.
format Preprint
id arxiv_https___arxiv_org_abs_2503_11662
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Lorecast: Layout-Aware Performance and Power Forecasting from Natural Language
Wang, Runzhi
Sengupta, Prianka
Roman-Vicharra, Cristhian
Chen, Yiran
Hu, Jiang
Hardware Architecture
Artificial Intelligence
Computation and Language
Machine Learning
In chip design planning, obtaining reliable performance and power forecasts for various design options is of critical importance. Traditionally, this involves using system-level models, which often lack accuracy, or trial synthesis, which is both labor-intensive and time-consuming. We introduce a new methodology, called Lorecast, which accepts English prompts as input to rapidly generate layout-aware performance and power estimates. This approach bypasses the need for HDL code development and synthesis, making it both fast and user-friendly. Experimental results demonstrate that Lorecast achieves accuracy within a few percent of error compared to post-layout analysis, while significantly reducing turnaround time.
title Lorecast: Layout-Aware Performance and Power Forecasting from Natural Language
topic Hardware Architecture
Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2503.11662