Zobrazeno 1 - 10
of 45
pro vyhledávání: '"David Defour"'
Autor:
David Defour
Publikováno v:
2022 IEEE 15th International Symposium on Embedded Multicore/Many-core Systems-on-Chip (MCSoC).
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::789e685af013e15239110ab3b347b292
https://doi.org/10.1109/mcsoc57363.2022.00040
https://doi.org/10.1109/mcsoc57363.2022.00040
Publikováno v:
ARITH
In this article, we propose to exploit the new computational capability offered by the Bfloat16 representation format to perform shadow computations and compute estimations of the relative error. We demonstrate and evaluate the assumptions under whic
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::bf9a64c9193b0ca2dac1af722b43c5f7
https://doi.org/10.1109/arith51176.2021.00017
https://doi.org/10.1109/arith51176.2021.00017
Autor:
Davide Mancusi, Alan Vaquet, Eric Petit, Matei Istoan, David Defour, Pablo de Oliveira Castro, Emeric Brun
Publikováno v:
ARITH
IEEE Transactions on Emerging Topics in Computing
IEEE Transactions on Emerging Topics in Computing, 2021, 9 (3), pp.1467-1478. ⟨10.1109/TETC.2021.3070422⟩
IEEE Transactions on Emerging Topics in Computing
IEEE Transactions on Emerging Topics in Computing, 2021, 9 (3), pp.1467-1478. ⟨10.1109/TETC.2021.3070422⟩
Mathematical libraries are typically developed for use with the fixed-width data-paths on processors and target common floating-point formats such as IEEE binary32 and binary64 . To address the increasing energy consumption and throughput requirement
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::18ed43b20a8a867a0bf0346b1fbecc53
https://doi.org/10.1109/arith51176.2021.00022
https://doi.org/10.1109/arith51176.2021.00022
Publikováno v:
IEEE 28th Symposium on Computer Arithmetic (ARITH)
IEEE 28th Symposium on Computer Arithmetic (ARITH), Jun 2021, Virtual Conference, France
HAL
IEEE 28th Symposium on Computer Arithmetic (ARITH), Jun 2021, Virtual Conference, France
HAL
International audience; In this article, we propose to exploit the new computational capability offered by the Bfloat16 representation format to perform shadow computations and compute estimations of the relative error. We demonstrate and evaluate th
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=dedup_wf_001::9d7668f1e251fd36be47f032e8341110
https://hal.archives-ouvertes.fr/hal-03159965/document
https://hal.archives-ouvertes.fr/hal-03159965/document
Publikováno v:
ARITH
[Research Report] Universite de Versailles; Intel corporation; Universite de Perpignan; Exascale Computing Research. 2020
[Research Report] Universite de Versailles; Intel corporation; Universite de Perpignan; Exascale Computing Research. 2020
The typical processors used for scientific computing have fixed-width data-paths. This implies that mathematical libraries were specifically developed to target each of these fixed precisions (binary16, binary 32, binary64). However, to address the i
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::24b36a872bf4587c2f406246262b1c72
https://doi.org/10.1109/arith48897.2020.00026
https://doi.org/10.1109/arith48897.2020.00026
Publikováno v:
IEEE Transactions on Computers
IEEE Transactions on Computers, Institute of Electrical and Electronics Engineers, 2017, 66 (12), pp.2058-2071. ⟨10.1109/TC.2017.2703870⟩
IEEE Transactions on Computers, Institute of Electrical and Electronics Engineers, 2017, 66 (12), pp.2058-2071. ⟨10.1109/TC.2017.2703870⟩
International audience; Elementary mathematical functions are pervasively used in many applications such as electronic calculators, computer simulations, or critical embedded systems. Their evaluation is always an approximation, which usually makes u
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::9eab1c3da04485436be5d18d2278c37a
https://doi.org/10.1109/tc.2017.2703870
https://doi.org/10.1109/tc.2017.2703870
Publikováno v:
Electric Power Systems Research
Electric Power Systems Research, Elsevier, 2017, 147, pp.81-87. ⟨10.1016/j.epsr.2017.02.017⟩
Electric Power Systems Research, Elsevier, 2017, 147, pp.81-87. ⟨10.1016/j.epsr.2017.02.017⟩
International audience; This paper presents a novel method based on midpoint-radius interval arithmetic to deal with uncertainties in the power ow problem. The proposed technique aims at nding a balance between accuracy and computational e ciency. It
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::66e17d2a76dbeb863db3452463f49c09
https://doi.org/10.1016/j.epsr.2017.02.017
https://doi.org/10.1016/j.epsr.2017.02.017
Publikováno v:
Euro-Par 2019: Parallel Processing
25th International Conference Euro-Par 2019 Parallel Processing
25th International Conference Euro-Par 2019 Parallel Processing, Aug 2019, Göttingen, Germany. pp.481-494, ⟨10.1007/978-3-030-29400-7_34⟩
Lecture Notes in Computer Science ISBN: 9783030293994
Euro-Par
25th International Conference Euro-Par 2019 Parallel Processing
25th International Conference Euro-Par 2019 Parallel Processing, Aug 2019, Göttingen, Germany. pp.481-494, ⟨10.1007/978-3-030-29400-7_34⟩
Lecture Notes in Computer Science ISBN: 9783030293994
Euro-Par
International audience; With the ever-increasing need for computation of scientific applications, new application domains, and major energy constraints, the landscape of floating-point computation is changing. New floating-point representation format
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::464f9076f25686bc8f0b0599063d467b
https://hal.archives-ouvertes.fr/hal-02564972/document
https://hal.archives-ouvertes.fr/hal-02564972/document
Autor:
Daniel Etiemble, David Defour
Publikováno v:
Technologies logicielles Architectures des systèmes.
Apres avoir rappele les differentes etapes qui ont mene des cartes graphiques des annees 1980 aux processeurs graphiques entierement programmables appeles GPU (2007), cet article presente les caracteristiques essentielles des GPU. La naissance de l
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::547d77787589fee70b5ffb98d73cf5f7
https://doi.org/10.51257/a-v1-h1013
https://doi.org/10.51257/a-v1-h1013
Autor:
Yohan Chatelain, Eric Petit, David Defour, Pablo de Oliveira Castro, Jordan Bieder, Marc Torrent
Publikováno v:
2018 IEEE 25th Symposium on Computer Arithmetic (ARITH)
2018 IEEE 25th Symposium on Computer Arithmetic (ARITH), Jun 2018, Amherst, United States. pp.61-68, ⟨10.1109/ARITH.2018.8464687⟩
ARITH
2018 IEEE 25th Symposium on Computer Arithmetic (ARITH), Jun 2018, Amherst, United States. pp.61-68, ⟨10.1109/ARITH.2018.8464687⟩
ARITH
VeriTracer automatically instruments a code and traces the accuracy of floating-point variables over time. VeriTracer enriches the visual traces with contextual information such as the call site path in which a value was modified. Contextual informat
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::336bfacf226bd9ddb5c4f36c66a8c3ce
https://hal.science/hal-01989607
https://hal.science/hal-01989607