Correct Rounding in Double Extended Precision (original) (raw)

Communication Dans Un Congrès Année : 2025

Résumé

The double extended precision format is an 80-bit floating-point format introduced in the 80x87 series of floating-point processors by Intel. Since the introduction of vector instructions in the x86 processors, its use has fallen due to speed concerns. We implement the first correctly-rounded routines for double extended precision. These implementations use modern microprocessor features and double-double arithmetic, avoiding x87-specific features, and achieve up to 2x speedup over state-of-the-art implementations which are not correctly rounded. This demonstrates that double extended precision could be viable as a large computational format.

Mots clés

• efficiency
• double extended format
• correct rounding
• floating-point
• IEEE 754
• IEEE 754 floating-point correct rounding double extended format efficiency

Dates et versions

hal-04861251 , version 1 (02-01-2025)

Licence

Identifiants

• HAL Id : hal-04861251 , version 1

Citer

Sélène Corbineau, Paul Zimmermann. Correct Rounding in Double Extended Precision. 32nd IEEE Symposium on Computer Arithmetic, May 2025, El Paso, TX, United States. ⟨hal-04861251⟩

631 Consultations

390 Téléchargements