Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 7
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Contribución de congreso (ConferenceOutput)

Anatomy of the BLIS family of algorithms for matrix multiplication

  • Castelló, Adrián|||0000-0002-8576-8451
  • Quintana-Ortí, Enrique S.|||0000-0002-5454-165X
  • Igual, Francisco D.
[EN] The efforts of the scientific community and hardware vendors to develop and optimize linear algebra codes have historically led to highly-tuned libraries, carefully adapted to the underlying processor architecture, with excellent (near-peak) performance. These optimization efforts, however, are commonly focused on obtaining the best performance possible when the involved operands are large and "squarish" matrices. New computationally-intensive applications (e.g., in deep learning) are increasingly demanding high-performance BLAS (Basic Linear Algebra Subprograms) also for small operands in any of their dimensions. In this paper, we tackle this problem by refactoring the general matrix-matrix multiplication (GEMM) algorithm within a specific high-performance implementation of BLAS, named BLIS, proposing a complete family of algorithmic variants to implement GEMM with different strategies to exploit the target cache hierarchy, together with the changes to be applied to architecture-specific codes to instantiate a complete GEMM implementation. Experimental results on an ARM processor (NVIDIA Carmel) reveal significant performance differences between the members of the GEMM family, depending on the shape and dimension of the matrix operands., This work was supported by the research project PID2020-113656RB-C22 of MCIN/AEI/10.13039/501100011033, EU (FEDER), the Spanish MINECO, under grants RTI2018-093684-B-I00, PID2021-126576NB-I00, and CM via Multiannual Agreement with Complutense University in the line Program to Stimulate Research for Young Doctors in the context of the V PRICIT under projects PR65/19-22445 and CM S2018/TCS-4423. A. Castello is a FJC2019-039222-I fellow supported by MCIN/AEI/10.13039/501100011033.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Contribución de congreso (ConferenceOutput)

Airborne ultrasonic technique to measure leaky Lamb waves with narrowband and low-frequency excitation

  • Vázquez, S.
  • Wright, W. M. D.
  • Gallardo, C.
  • Gosálbez Castillo, Jorge|||0000-0001-6520-9014
  • Bosch Roig, Ignacio|||0000-0003-3190-3635
  • CARRIÓN GARCÍA, ALICIA|||0000-0002-0630-6065
Air-coupled non-destructive testing has become very valuable because eliminates the use of liquid couplants and enables a contactless fast inspection of structures. Multimodal leaky Lamb waves can be excited and sensed with this technology in plate-like specimens varying the excitation frequency and the inclination of the transducers. To reduce undesired effects, as the existence of unwanted modes, narrowband and low-frequency excitation signals can be used. In this paper, the generation and detection of ultrasonic leaky Lamb waves in a metallic plate with capacitive air-coupled transducers have been performed. Theoretical and experimental data have been matched, showing good correlation., The authors acknowledge the support from University College Cork (Ireland), Universitat Politècnica de València and the Spanish Administration under grants BIA2017-87573 and Salvador Madariaga's Programme (PR2017-00658).
Repositorio Institucional de la Universidad de Burgos (RIUBU)
Publication Contribución de congreso (ConferenceOutput)

The urban freight distribution in medium size cities: Descriptive data taken from Pamplona (Spain) and Angers (France)

  • Serrano Hernández, Adrián
  • Gougeon, Thomas
  • Cadarso, Luis
  • Juan, Angel A.
  • Faulín Fajardo, Francisco Javier
Trabajo presentado en: R-Evolucionando el transporte, XIV Congreso de Ingeniería del Transporte (CIT 2021), realizado en modalidad online los días 6, 7 y 8 de julio de 2021, organizado por la Universidad de Burgos, Medium-size cities around the world are experiencing rapid changes in their urban centers regarding sustainable mobility, pedestrianization of streets, and vehicle-access control to their city center. This situation is particularly important in the last-mile urban distribution in city centers due to the fact that private cars, public transportation, and freight transportation fiery compete for the same space. This article contextualizes this phenomenon in two European medium-size cities: Angers (France) and Pamplona (Spain). The paper also draws the current situation regarding their mobility and freight-delivery systems in their city centers. Thus, an apposite survey has been deployed in both cities in order to collect social perceptions and mobility patterns data. Additionally, we have included a conjoint analysis study for examining the characteristic determinants for aerial (drone) distribution. Likewise, meaningful conclusions can be extracted from the analysis of the results generated by the current analysis. Firstly, we have realized the immediate effects of a recent sustainable urban-mobility plan at Pamplona, which contrasts with the Angers case, where no similar plan has been deployed yet. Secondly, we have focused our attention on the socio-economic factors that determine the mobility in both cities. Thirdly, we have observed, in both cities, a clear preference for electric vehicles and cargo-bikes for driving inside the city center. Finally, there is a rejection to use drones for merchandise delivery., We would like to acknowledge the partial financial support given by the Spanish Ministry of Science, Innovation, and Universities (RED2018-102642-T), and the following projects: IoF2020-H2020 (731884), “la Caixa” Foundation (LCF/PR/PR15/51100007), and AGAUR (2018-LLAV-00017) to the authors of this paper.
Docta Complutense
Publication Contribución de congreso (ConferenceOutput)

Tools for the use of SignWriting as a language resource

  • García Sevilla, Antonio Fernando
  • Díaz Esteban, Alberto
  • Lahoz Bengoechea, José María
Representation of linguistic data is an issue of utmost importance when developing language resources, but the lack of a standard written form in sign languages presents a challenge. Different notation systems exist, but only SignWriting seems to have some use in the native signer community. It is, however, a difficult system to use computationally, not based on a linear sequence of characters. We present the project “VisSE”, which aims to develop tools for the effective use of SignWriting in the computer. The first of these is an application which uses computer vision to interpret SignWriting, understanding the meaning of new or existing transcriptions, or even hand-written images. Two additional tools will be able to consume the result of this recognizer: first, a textual description of the features of the transcription will make it understandable for non-signers. Second, a three-dimensional avatar will be able to reproduce the configurations and movements contained within the transcription, making it understandable for signers even if not familiar with SignWriting. Additionally, the project will result in a corpus of annotated SignWriting data which will also be of use to the computational linguistics community.
Docta Complutense
Publication Contribución de congreso (ConferenceOutput)

Quevedo: Annotation and processing of graphical languages

  • García Sevilla, Antonio Fernando
  • Díaz Esteban, Alberto
  • Lahoz Bengoechea, José María
In this article, we present Quevedo, a software tool we have developed for the task of automatic processing of graphical languages. These are languages which use images to convey meaning, relying not only on the shape of symbols but also on their spatial arrangement in the page, and relative to each other. When presented in image form, these languages require specialized computational processing which is not the same as usually done either for natural language processing or for artificial vision. Quevedo enables this specialized processing, focusing on a data-based approach. As a command line application and library, it provides features for the collection and management of image datasets, and their machine learning recognition using neural networks and recognizer pipelines. This processing requires careful annotation of the source data, for which Quevedo offers an extensive and visual web-based annotation interface. In this article, we also briefly present a case study centered on the task of SignWriting recognition, the original motivation for writing the software. Quevedo is written in Python, and distributed freely under the Open Software License version 3.0.
Academica-e. Repositorio Institucional de la Universidad Pública de Navarra
Publication Contribución de congreso (ConferenceOutput)

Medida de torsión en seguidores solares con sensores de fibra óptica, Torsion measurements on solar trackers using fiber optic sensors

  • Leandro González, Daniel
  • Bravo Acha, Mikel
  • López-Amo Sáinz, Manuel
  • Loayssa Lara, Alayn
  • Júdez Colorado, Aitor
  • Mariñelarena Ollacarizqueta, Jon
  • Falcone Lanas, Francisco
  • Jiménez Romero, Sergio
  • Moriana García, Iñaki
En este trabajo se ha monitorizado la torsión sufrida por una estructura de seguidores solares empleando sensores de fibra óptica. Estas medidas de campo se han obtenido con una red de sensores formada por 45 redes de difracción de Bragg en una huerta solar activa. Estos resultados, correlados con la velocidad y dirección del viento permitirían aumentar la eficiencia de los seguidores y ayudar a entender y prevenir efectos negativos relacionados con inestabilidades aeroelásticas., In this work, the torsion experienced by set of solar trackers is monitored using fiber optic sensors. The in-field measurements were obtained using a sensor network comprising 45 fiber Bragg gratings in an active solar plant. These results, correlated with wind speed and direction could increase efficiency and lead to understand and prevent harmful effects related to aeroelastic instability events., Este trabajo ha sido financiado por el proyecto WINDSOLAR (0011-1365-2017-000122) del Gobierno de Navarra, los proyectos TEC2016-76021-C2 y PID2019-107270RB-C22, los fondos FEDER, la European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 838143 y las fundaciones 'Caja Navarra' y 'la Caixa' con el proyecto LCF/PR/PR13/51080004.
Academica-e. Repositorio Institucional de la Universidad Pública de Navarra
Publication Contribución de congreso (ConferenceOutput)

Sistema de interrogación multi-sensor basado en multiplexación en longitud de onda y tiempo, Multi-sensor interrogation system based on wavelength and time multiplexing

  • Bravo Acha, Mikel
  • Leandro González, Daniel
  • Pérez González, Asier
  • López-Amo Sáinz, Manuel
En este trabajo se propone un sistema de interrogación versátil y económico basada en la generación de pulsos empleando la modulación directa de diodos láser DFB. Se ha demostrado una resolución en longitud de onda de 1 pm para una separación de sensores de 600m. La validación del interrogador se ha llevado a cabo monitorizando una red de sensores de 26 km de longitud que incluye redes de difracción de Bragg, reflectores y un interferómetro Fabry-Pérot., This work proposes a versatile and cost-effective interrogation system based on pulses generation using direct modulated DFB laser diodes. A wavelength resolution of 1 pm is demonstrated for a 600 m sensor-spacing. A 26 km-long sensor network including fiber Bragg gratings, wideband reflectors and a Fabry-Pérot interferometer is used to validate the interrogator., Este trabajo ha sido financiado por los proyectos TEC2016-76021-C2 y PID2019-107270RB-C22, los fondos FEDER, la European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 838143, el contrato Beatriz Galindo BEAGAL18/00116 y las fundaciones 'Caja Navarra' y 'la Caixa' con el proyecto LCF/PR/PR13/51080004.