Automatic Tunable Floating-Point Format Selection in a RISC-V Ecosystem

Alberto Nannarelli; Marco Re; Gian Carlo Cardarilli; Luca Di Nunzio; Sergio Spanò

doi:10.1007/978-3-031-84100-2_8

Automatic Tunable Floating-Point Format Selection in a RISC-V Ecosystem

Alberto Nannarelli, Marco Re^*, Gian Carlo Cardarilli, Luca Di Nunzio, Sergio Spanò

^*Corresponding author for this work

University of Rome Tor Vergata

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

In this work, we address variable precision computation that can be adjusted depending on the characteristics and the specification of the algorithm. We base our method on the Tunable Floating-Point (TFP) format, which allows to change format in floating-point units to tune the precision and the dynamic range to specific parts of the algorithm. We propose an automatic method to change the TFP format during the computation, based on a quality function, which is embedded in the execution of the arithmetic kernel.

Original language	English
Title of host publication	Proceedings of the International Conference on Applications in Electronics Pervading Industry, Environment and Society - ApplePies 2024
Volume	1369
Publisher	Springer
Publication date	2025
Pages	65-73
ISBN (Print)	978-3-031-84099-9
ISBN (Electronic)	978-3-031-84100-2
DOIs	https://doi.org/10.1007/978-3-031-84100-2_8
Publication status	Published - 2025
Event	International Conference on Applications in Electronics Pervading Industry, Environment and Society 2024 - Turin, Italy Duration: 19 Sept 2024 → 20 Sept 2024

Conference

Conference	International Conference on Applications in Electronics Pervading Industry, Environment and Society 2024
Country/Territory	Italy
City	Turin
Period	19/09/2024 → 20/09/2024

Access to Document

10.1007/978-3-031-84100-2_8

OpenUrl availability

Full text

Cite this

Nannarelli, A., Re, M., Cardarilli, G. C., Di Nunzio, L., & Spanò, S. (2025). Automatic Tunable Floating-Point Format Selection in a RISC-V Ecosystem. In Proceedings of the International Conference on Applications in Electronics Pervading Industry, Environment and Society - ApplePies 2024 (Vol. 1369, pp. 65-73). Springer. https://doi.org/10.1007/978-3-031-84100-2_8

@inproceedings{f2fd53f64f5b4d27b8b988f4c09ea243,

title = "Automatic Tunable Floating-Point Format Selection in a RISC-V Ecosystem",

abstract = "In this work, we address variable precision computation that can be adjusted depending on the characteristics and the specification of the algorithm. We base our method on the Tunable Floating-Point (TFP) format, which allows to change format in floating-point units to tune the precision and the dynamic range to specific parts of the algorithm. We propose an automatic method to change the TFP format during the computation, based on a quality function, which is embedded in the execution of the arithmetic kernel.",

author = "Alberto Nannarelli and Marco Re and Cardarilli, {Gian Carlo} and {Di Nunzio}, Luca and Sergio Span{\`o}",

year = "2025",

doi = "10.1007/978-3-031-84100-2_8",

language = "English",

isbn = "978-3-031-84099-9",

volume = "1369",

pages = "65--73",

booktitle = "Proceedings of the International Conference on Applications in Electronics Pervading Industry, Environment and Society - ApplePies 2024",

publisher = "Springer",

note = "International Conference on Applications in Electronics Pervading Industry, Environment and Society 2024, ApplePies 2024 ; Conference date: 19-09-2024 Through 20-09-2024",

}

Nannarelli, A, Re, M, Cardarilli, GC, Di Nunzio, L & Spanò, S 2025, Automatic Tunable Floating-Point Format Selection in a RISC-V Ecosystem. in Proceedings of the International Conference on Applications in Electronics Pervading Industry, Environment and Society - ApplePies 2024. vol. 1369, Springer, pp. 65-73, International Conference on Applications in Electronics Pervading Industry, Environment and Society 2024, Turin, Italy, 19/09/2024. https://doi.org/10.1007/978-3-031-84100-2_8

Automatic Tunable Floating-Point Format Selection in a RISC-V Ecosystem. / Nannarelli, Alberto; Re, Marco; Cardarilli, Gian Carlo et al.
Proceedings of the International Conference on Applications in Electronics Pervading Industry, Environment and Society - ApplePies 2024. Vol. 1369 Springer, 2025. p. 65-73.

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - Automatic Tunable Floating-Point Format Selection in a RISC-V Ecosystem

AU - Nannarelli, Alberto

AU - Re, Marco

AU - Cardarilli, Gian Carlo

AU - Di Nunzio, Luca

AU - Spanò, Sergio

PY - 2025

Y1 - 2025

N2 - In this work, we address variable precision computation that can be adjusted depending on the characteristics and the specification of the algorithm. We base our method on the Tunable Floating-Point (TFP) format, which allows to change format in floating-point units to tune the precision and the dynamic range to specific parts of the algorithm. We propose an automatic method to change the TFP format during the computation, based on a quality function, which is embedded in the execution of the arithmetic kernel.

AB - In this work, we address variable precision computation that can be adjusted depending on the characteristics and the specification of the algorithm. We base our method on the Tunable Floating-Point (TFP) format, which allows to change format in floating-point units to tune the precision and the dynamic range to specific parts of the algorithm. We propose an automatic method to change the TFP format during the computation, based on a quality function, which is embedded in the execution of the arithmetic kernel.

U2 - 10.1007/978-3-031-84100-2_8

DO - 10.1007/978-3-031-84100-2_8

M3 - Article in proceedings

SN - 978-3-031-84099-9

VL - 1369

SP - 65

EP - 73

BT - Proceedings of the International Conference on Applications in Electronics Pervading Industry, Environment and Society - ApplePies 2024

PB - Springer

T2 - International Conference on Applications in Electronics Pervading Industry, Environment and Society 2024

Y2 - 19 September 2024 through 20 September 2024