Microcontroller (MCU) performance screening ensures that devices meet critical specifications, such as maximum operating frequency (Fmax). On-chip Speed Monitors (SMONs), implemented as ring oscillators, provide process-correlated signals that can be used to estimate Fmax via machine learning (ML). However, traditional ML models require substantial domain expertise, extensive feature engineering, hyperparameter tuning, and dataset-specific training, limiting their scalability and generalization. In this preliminary study, we explore the use of TabPFN, a pre-trained Tabular Foundation Model (TabFM) based on In-Context Learning (ICL), for MCU performance prediction. TabPFN eliminates the need for task-specific training or tuning by conditioning directly on labeled examples provided at inference time, enabling few-shot and zero-shot learning. We evaluate TabPFN on two distinct MCU datasets and compare its performance with conventional ML models, including tree-based and linear approaches. Our results show that TabPFN consistently achieves competitive accuracy with minimal human supervision, demonstrating its potential as a fast, generalizable, and low-maintenance alternative for performance screening in semiconductor manufacturing.

In-Context Learning for Microcontroller Performance Screening Using Tabular Foundation Models / Bellarmino, Nicolò; Cantoro, Riccardo; Huch, Martin; Kilian, Tobias; Squillero, Giovanni. - ELETTRONICO. - (In corso di stampa). (Intervento presentato al convegno 28th Euromicro Conference Series on Digital System Design (DSD) 2025 tenutosi a Salerno (IT) nel 10-12 September, 2025).

In-Context Learning for Microcontroller Performance Screening Using Tabular Foundation Models

Nicolò Bellarmino;Riccardo Cantoro;Giovanni Squillero
In corso di stampa

Abstract

Microcontroller (MCU) performance screening ensures that devices meet critical specifications, such as maximum operating frequency (Fmax). On-chip Speed Monitors (SMONs), implemented as ring oscillators, provide process-correlated signals that can be used to estimate Fmax via machine learning (ML). However, traditional ML models require substantial domain expertise, extensive feature engineering, hyperparameter tuning, and dataset-specific training, limiting their scalability and generalization. In this preliminary study, we explore the use of TabPFN, a pre-trained Tabular Foundation Model (TabFM) based on In-Context Learning (ICL), for MCU performance prediction. TabPFN eliminates the need for task-specific training or tuning by conditioning directly on labeled examples provided at inference time, enabling few-shot and zero-shot learning. We evaluate TabPFN on two distinct MCU datasets and compare its performance with conventional ML models, including tree-based and linear approaches. Our results show that TabPFN consistently achieves competitive accuracy with minimal human supervision, demonstrating its potential as a fast, generalizable, and low-maintenance alternative for performance screening in semiconductor manufacturing.
In corso di stampa
File in questo prodotto:
File Dimensione Formato  
2025_DSD_Foundation_Models (2).pdf

accesso aperto

Tipologia: 2. Post-print / Author's Accepted Manuscript
Licenza: Pubblico - Tutti i diritti riservati
Dimensione 472.62 kB
Formato Adobe PDF
472.62 kB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11583/3002058