Suchergebnisse - "[INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic"

Correctly-Rounded Evaluation of a Function: Why, How, and at What Cost?

Autoren: Brisebarre, Nicolas Hanrot, Guillaume Muller, Jean-Michel et al.

Weitere Verfasser: Brisebarre, Nicolas Hanrot, Guillaume Muller, Jean-Michel et al.

Quelle: ACM Computing Surveys. 58:1-34

Schlagwörter: algorithmic number theory, standardization, elementary functions, floating-point arithmetic, LLL algorithm, lattice basis reduction, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, approximation theory, Computer arithmetic, Correct rounding, Table maker's dilemma

Dateibeschreibung: application/pdf

Truncated QR Factorization with Pivoting in Mixed Precision: Truncated QR factorization with pivoting in mixed precision

Autoren: Buttari, Alfredo Mary, Theo Pacteau, André

Weitere Verfasser: Buttari, Alfredo Mary, Theo Pacteau, André

Quelle: SIAM Journal on Scientific Computing. 47:B382-B401

Schlagwörter: mixed precision, [INFO.INFO-MS] Computer Science [cs]/Mathematical Software [cs.MS], Roundoff error, [INFO.INFO-NA] Computer Science [cs]/Numerical Analysis [cs.NA], Mixed-precision algorithms, low-rank approximations, QR factorization, Numerical methods for low-rank matrix approximation, matrix compression, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic

Dateibeschreibung: application/xml; application/pdf

Emulation of the FMA and the correctly-rounded sum of three numbers in rounding-to-nearest floating-point arithmetic

Autoren: Graillat, Stef Muller, Jean-Michel

Weitere Verfasser: Graillat, Stef Muller, Jean-Michel

Quelle: Numerische Mathematik.

Schlagwörter: triple-word arithmetic, fused multiply-add, double-word arithmetic, Error-free transforms, FMA, Floating-point arithmetic, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, ADD3

Dateibeschreibung: application/pdf

Xvpfloat: RISC-V ISA Extension for Variable Extended Precision Floating Point Computation

Autoren: Guthmuller, Eric Fuguet, César Bocco, Andrea et al.

Weitere Verfasser: Guthmuller, Eric Fuguet, César Bocco, Andrea et al.

Quelle: IEEE Transactions on Computers

Schlagwörter: [INFO.INFO-AR]Computer Science [cs]/Hardware Architecture [cs.AR], [INFO.INFO-AR] Computer Science [cs]/Hardware Architecture [cs.AR], [INFO.INFO-AO]Computer Science [cs]/Computer Arithmetic, [INFO.INFO-NA] Computer Science [cs]/Numerical Analysis [cs.NA], [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, [INFO.INFO-NA]Computer Science [cs]/Numerical Analysis [cs.NA], 7. Clean energy

Dateibeschreibung: application/pdf

Zugangs-URL: https://cea.hal.science/cea-04546949v1
https://doi.org/10.1109/tc.2024.3383964
https://cea.hal.science/cea-04546949v1/document

Rounding Error Analysis of an Orbital Collision Probability Evaluation Algorithm

Autoren: Arzelier, Denis Bréhard, Florent Joldes, Mioara et al.

Weitere Verfasser: Arzelier, Denis Bréhard, Florent Joldes, Mioara et al.

Quelle: 2024 IEEE 31st Symposium on Computer Arithmetic (ARITH). :96-103

Schlagwörter: D-finite function, Evaluation error, Error bound, Orbital collision probability, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, Floating point arithmetic, Majorizing series

Dateibeschreibung: application/pdf

An Emacs-Cairo Scrolling Bug due to Floating-Point Inaccuracy

Autoren: Lefèvre, Vincent

Weitere Verfasser: Lefèvre, Vincent

Quelle: 2024 IEEE 31st Symposium on Computer Arithmetic (ARITH). :76-79

Schlagwörter: rounding, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, floating-point, discontinuity

Dateibeschreibung: application/pdf

Zugangs-URL: https://inria.hal.science/hal-04566768v1
https://inria.hal.science/hal-04566768v1/document
https://doi.org/10.1109/arith61463.2024.00022

Useful applications of correctly-rounded operators of the form ab + cd + e

Autoren: Hubrecht, Tom Jeannerod, Claude-Pierre Muller, Jean-Michel

Weitere Verfasser: Hubrecht, Tom Jeannerod, Claude-Pierre Muller, Jean-Michel

Quelle: 2024 IEEE 31st Symposium on Computer Arithmetic (ARITH). :32-39

Schlagwörter: floating-point arithmetic, fused operators, [INFO.INFO-MS] Computer Science [cs]/Mathematical Software [cs.MS], double-word arithmetic, [INFO.INFO-NA] Computer Science [cs]/Numerical Analysis [cs.NA], [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, correct rounding, complex arithmetic, [MATH.MATH-NA] Mathematics [math]/Numerical Analysis [math.NA], dot-product instructions, rounding error analysis

Dateibeschreibung: application/pdf

Zugangs-URL: https://inria.hal.science/hal-04461089v2

AdaQAT: Adaptive Bit-Width Quantization-Aware Training

Autoren: Gernigon, Cédric Filip, Silviu-Ioan Sentieys, Olivier et al.

Weitere Verfasser: Gernigon, Cédric Filip, Silviu-Ioan Sentieys, Olivier et al.

Quelle: 2024 IEEE 6th International Conference on AI Circuits and Systems (AICAS). :442-446

Schlagwörter: [INFO.INFO-AI] Computer Science [cs]/Artificial Intelligence [cs.AI], FOS: Computer and information sciences, Computer Science - Machine Learning, Artificial Intelligence (cs.AI), Computer Science - Artificial Intelligence, Adaptive Bit-Width Optimization, Neural Network Compression, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, Quantization Aware Training, Machine Learning (cs.LG)

Dateibeschreibung: application/pdf

Zugangs-URL: http://arxiv.org/abs/2404.16876
https://hal.science/hal-04549245v1

Efficient and Validated Numerical Evaluation of Abelian Integrals

Autoren: Florent Bréhard Nicolas Brisebarre Mioara Joldes et al.

Weitere Verfasser: Florent Bréhard Nicolas Brisebarre Mioara Joldes et al.

Quelle: ACM Transactions on Mathematical Software. 50:1-38

Schlagwörter: [INFO.INFO-SC]Computer Science [cs]/Symbolic Computation [cs.SC], limit cycles, [INFO.INFO-AO]Computer Science [cs]/Computer Arithmetic, [INFO.INFO-SC] Computer Science [cs]/Symbolic Computation [cs.SC], [MATH.MATH-DS]Mathematics [math]/Dynamical Systems [math.DS], Abelian integral, Newton-like operator, [MATH.MATH-DS] Mathematics [math]/Dynamical Systems [math.DS], 37M20, rigorous numerics, Hilbert's 16th problem, [INFO.INFO-NA]Computer Science [cs]/Numerical Analysis [cs.NA], Secondary: 37G15, 01 natural sciences, 2020 AMS classification: Primary: 34C07, [INFO.INFO-NA] Computer Science [cs]/Numerical Analysis [cs.NA], [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, 42A10, trigonometric polynomial interpolation, 0101 mathematics, 42A15, 65G20

Dateibeschreibung: application/pdf

Zugangs-URL: https://hal.science/hal-03561096v2/document
https://hal.science/hal-03561096v2
https://doi.org/10.1145/3637550

Five-Precision GMRES-Based Iterative Refinement

Autoren: Amestoy, Patrick Buttari, Alfredo Higham, Nicholas et al.

Weitere Verfasser: Amestoy, Patrick Buttari, Alfredo Higham, Nicholas et al.

Quelle: Amestoy, P, Buttari, A, Higham, N J, L’EXCELLENT, JEAN-YVES, Mary, T & Vieuble, B 2023, 'FIVE-PRECISION GMRES-BASED ITERATIVE REFINEMENT ∗', SIAM Journal on Matrix Analysis and Applications.

Schlagwörter: mixed precision, floating-point arithmetic, [INFO.INFO-AO]Computer Science [cs]/Computer Arithmetic, 4. Education, [INFO.INFO-NA]Computer Science [cs]/Numerical Analysis [cs.NA], GMRES, rounding error analysis, [INFO.INFO-MS] Computer Science [cs]/Mathematical Software [cs.MS], preconditioning, [INFO.INFO-NA] Computer Science [cs]/Numerical Analysis [cs.NA], iterative refinement, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, linear system, multiple precision, 65G50, backward error, forward error, [INFO.INFO-MS]Computer Science [cs]/Mathematical Software [cs.MS]

Dateibeschreibung: application/pdf

Zugangs-URL: https://research.manchester.ac.uk/en/publications/f377b09f-1e32-4247-8494-0c69c7763410
https://pure.manchester.ac.uk/ws/files/281551091/paper.pdf
https://research.manchester.ac.uk/en/publications/f377b09f-1e32-4247-8494-0c69c7763410

Towards Multi-Level Arithmetic Optimizations

Autoren: Cochard, Pierre Forget, Luc de Dinechin, Florent et al.

Weitere Verfasser: Cochard, Pierre Forget, Luc de Dinechin, Florent et al.

Schlagwörter: [INFO.INFO-AI] Computer Science [cs]/Artificial Intelligence [cs.AI], computer arithmetic, [INFO.INFO-AR] Computer Science [cs]/Hardware Architecture [cs.AR], flopoco, mlir, llvm, [INFO.INFO-CL] Computer Science [cs]/Computation and Language [cs.CL], hardare design, compilation, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, high level synthesis, [INFO] Computer Science [cs]

Dateibeschreibung: application/pdf

Zugangs-URL: https://hal.science/hal-05063466v1

Modular Multiplication in the AMNS Representation: Hardware Implementation

Autoren: Noyez, Louis El Mrabet, Nadia Potin, Olivier et al.

Weitere Verfasser: Noyez, Louis El Mrabet, Nadia Potin, Olivier et al.

Quelle: Lecture Notes in Computer Science ISBN: 9783031828409

Schlagwörter: [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, AMNS, AMNS Montgomery multiplication Hardware Implementation, Hardware Implementation, Montgomery multiplication

Dateibeschreibung: application/pdf

Zugangs-URL: https://hal.science/hal-04691484v1

A Stochastic Rounding-Enabled Low-Precision Floating-Point MAC for DNN Training

Autoren: Ben Ali, Sami Filip, Silviu-Ioan Sentieys, Olivier et al.

Weitere Verfasser: Ben Ali, Sami Filip, Silviu-Ioan Sentieys, Olivier et al.

Quelle: 2024 Design, Automation & Test in Europe Conference & Exhibition (DATE). :1-6

Schlagwörter: FOS: Computer and information sciences, 0301 basic medicine, [INFO.INFO-AR] Computer Science [cs]/Hardware Architecture [cs.AR], 0303 health sciences, 03 medical and health sciences, Hardware Architecture (cs.AR), stochastic rounding, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, low-precision MAC unit, Computer Science - Hardware Architecture, DNN

Dateibeschreibung: application/pdf

Zugangs-URL: http://arxiv.org/abs/2404.14010
https://hal.science/hal-04380270v3
https://hal.science/hal-04380270v3/document

Performance of precision auto-tuned neural networks

Autoren: Ferro, Quentin Graillat, Stef Hilaire, Thibault et al.

Weitere Verfasser: Ferro, Quentin Graillat, Stef Hilaire, Thibault et al.

Quelle: 2023 IEEE 16th International Symposium on Embedded Multicore/Many-core Systems-on-Chip (MCSoC). :592-599

Schlagwörter: Neural Networks, [INFO.INFO-PF] Computer Science [cs]/Performance [cs.PF], [INFO.INFO-NA] Computer Science [cs]/Numerical Analysis [cs.NA], [INFO.INFO-NE] Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], Stochastic Arithmetic, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, Auto-Tuning, Precision, Floating-Point

Dateibeschreibung: application/pdf

Zugangs-URL: https://hal.science/hal-04149501v1

Low-Precision Floating-Point for Efficient On-Board Deep Neural Network Processing

Autoren: Gernigon, Cédric Filip, Silviu-Ioan Sentieys, Olivier et al.

Weitere Verfasser: Gernigon, Cédric Filip, Silviu-Ioan Sentieys, Olivier et al.

Quelle: 2023 European Data Handling & Data Processing Conference (EDHPC). :1-8

Schlagwörter: [INFO.INFO-AI] Computer Science [cs]/Artificial Intelligence [cs.AI], Signal Processing (eess.SP), FOS: Computer and information sciences, [INFO.INFO-AR] Computer Science [cs]/Hardware Architecture [cs.AR], Computer Science - Machine Learning, Ship Detection, Deep Neural Networks (DNN), Machine Learning (cs.LG), FOS: Electrical engineering, electronic engineering, information engineering, Semantic Segmentation, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, Reduced Precision, Electrical Engineering and Systems Science - Signal Processing, Quantization-Aware Training (QAT), Floating-Point

Dateibeschreibung: application/pdf

Zugangs-URL: http://arxiv.org/abs/2311.11172
https://hal.science/hal-04252197v1/document
https://doi.org/10.23919/edhpc59100.2023.10396014
https://hal.science/hal-04252197v1

Variable and Extended Precision (VRP) Accelerator Implemented in a 22 nm SoC

Autoren: Guthmuller, Eric Fuguet, César Bocco, Andrea et al.

Weitere Verfasser: Guthmuller, Eric Fuguet, César Bocco, Andrea et al.

Quelle: Electronics Letters. 61

Schlagwörter: [INFO.INFO-AR] Computer Science [cs]/Hardware Architecture [cs.AR], [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic

Dateibeschreibung: application/pdf

Zugangs-URL: https://cea.hal.science/cea-05123539v1
https://cea.hal.science/cea-05123539v1/document
https://doi.org/10.1049/ell2.70255

On Polynomial Modular Number Systems over $ \mathbb{Z}/{p}\mathbb{Z} $

Autoren: Bajard, Jean-Claude Marrez, Jérémy Plantard, Thomas et al.

Weitere Verfasser: Bajard, Jean-Claude Marrez, Jérémy Plantard, Thomas et al.

Quelle: Advances in Mathematics of Communications. 18:674-695

Schlagwörter: FOS: Computer and information sciences, Mathematics - Number Theory, Polynomial roots, Computer Science - Data Structures and Algorithms, Modular Computation, FOS: Mathematics, Finite field, [INFO.INFO-AO] Computer Science [cs]/Computer Arithmetic, Data Structures and Algorithms (cs.DS), Number Theory (math.NT), 0101 mathematics, 01 natural sciences, Polynomial Modular Number System

Dateibeschreibung: application/pdf

Zugangs-URL: http://arxiv.org/abs/2001.03741
https://hal.science/hal-03611829v1
https://hal.science/hal-03611829v1/document
https://doi.org/10.3934/amc.2022018

Hardware Level Simulations of Fault Detection in RNS Accelerator

Autoren: Vollmer, Morgane Tisserand, Arnaud Bigou, Karim et al.

Weitere Verfasser: Vollmer, Morgane Tisserand, Arnaud Bigou, Karim et al.

Quelle: BITFLIP 2025 ; https://hal.science/hal-05368416 ; BITFLIP 2025, Nov 2025, Rennes, France. ; https://www.creachlabs.fr/fr/les-conferences-des-rencontres-scientifiques-enjeux/bitflip-2025-cyber-reliability-and-tolerance-faults-electronic-components

Schlagwörter: [INFO.INFO-AO]Computer Science [cs]/Computer Arithmetic, [INFO.INFO-AR]Computer Science [cs]/Hardware Architecture [cs.AR], [INFO.INFO-CR]Computer Science [cs]/Cryptography and Security [cs.CR]

Geographisches Schlagwort: Rennes, France

Verfügbarkeit: https://hal.science/hal-05368416
https://hal.science/hal-05368416v1/document
https://hal.science/hal-05368416v1/file/poster.pdf

Electrical Activity in Arithmetic Operations vs Security

Autoren: Tisserand, Arnaud Centre National de la Recherche Scientifique (CNRS) Laboratoire des sciences et techniques de l'information, de la communication et de la connaissance (Lab-STICC) et al.

Weitere Verfasser: Tisserand, Arnaud Centre National de la Recherche Scientifique (CNRS) Laboratoire des sciences et techniques de l'information, de la communication et de la connaissance (Lab-STICC) et al.

Quelle: RAIM Meeting 2025: 16th Rencontres de l'Arithmétique en Informatique Mathématique ; https://hal.science/hal-05347564 ; RAIM Meeting 2025: 16th Rencontres de l'Arithmétique en Informatique Mathématique, Nov 2025, Lyon, France ; https://raim2025.sciencesconf.org/

Schlagwörter: [INFO.INFO-AO]Computer Science [cs]/Computer Arithmetic, [INFO.INFO-AR]Computer Science [cs]/Hardware Architecture [cs.AR], [INFO.INFO-CR]Computer Science [cs]/Cryptography and Security [cs.CR]

Geographisches Schlagwort: Lyon, France

Verfügbarkeit: https://hal.science/hal-05347564

SERA-Float: A Soft Error Resilient Approximate Floating-Point Computing Format

Autoren: Mishra, Vishesh Traiola, Marcello Kritikakou, Angeliki et al.

Weitere Verfasser: Mishra, Vishesh Traiola, Marcello Kritikakou, Angeliki et al.

Quelle: ICCAD 2025 - ACM/IEEE International Conference On Computer Aided Design ; https://inria.hal.science/hal-05333255 ; ICCAD 2025 - ACM/IEEE International Conference On Computer Aided Design, Oct 2025, Munich (Allemagne), Germany. pp.1-9

Schlagwörter: Neural Networks, Fault-Tolerant Computing, Energy Efficiency, Floating-point Format, Approximate Computing, [INFO]Computer Science [cs], [INFO.INFO-AO]Computer Science [cs]/Computer Arithmetic, [INFO.INFO-AR]Computer Science [cs]/Hardware Architecture [cs.AR], [INFO.INFO-ES]Computer Science [cs]/Embedded Systems

Geographisches Schlagwort: Munich (Allemagne), Germany

Verfügbarkeit: https://inria.hal.science/hal-05333255
https://inria.hal.science/hal-05333255v1/document
https://inria.hal.science/hal-05333255v1/file/SERA_FLOAT_v2.pdf