ARCO - Microarquitectura i Compiladors
Les activitats de recerca del grup ARCO es centren en l'area de arquitectura de computadors, compiladors i el processament en paral·lel, amb especial èmfasi en la microarquitectura i les tècniques de generació de codi per a sistemes de computació fiables i eficients energèticament. Un dels seus principals enfocaments actuals és sobre sistemes de computació intel·ligents, on l'objectiu és dissenyar noves arquitectures per a l'aprenentatge automàtic, la visió per computador i el processament del llenguatge. L'altre enfocament principal és en els processadors gràfics tant per a càrregues de treball de propòsit general com per a aplicacions gràfiques.
El grup està format per professors i estudiants de la Universitat Politècnica de Catalunya, la Universitat de Múrcia i la Universitat Rovira i Virgili. El grup té un llarg historial de publicacions científiques, amb més de 500 articles d'investigació, i transferències de tecnologia, amb més de 50 patents.
Las actividades de investigación del grupo ARCO se centran en el área de arquitectura de computadores, compiladores y el procesamiento en paralelo, con especial énfasis en la microarquitectura y las técnicas de generación de código para sistemas de computación fiables y eficientes energéticamente. Uno de sus principales enfoques actuales es sobre sistemas de computación inteligentes, donde el objetivo es diseñar nuevas arquitecturas para el aprendizaje automático, la visión por computador y el procesamiento del lenguaje. El otro enfoque principal es en los procesadores gráficos tanto para cargas de trabajo de propósito general como para aplicaciones gráficas.
El grupo está formado por profesores y estudiantes de la Universidad Politécnica de Catalunya, la Universidad de Murcia y la Universidad Rovira i Virgili. El grupo tiene un largo historial de publicaciones científicas, con más de 500 artículos de investigación, y transferencias de tecnología, con más de 50 patentes.
The research activities of the ARCO group focus on computer architecture, compilers and parallel processing, with special emphasis on microarchitecture and code generation techniques for energy-efficient and reliable computing systems. One of its main current focuses is on intelligent computing systems, where the goal is to devise novel architectures for machine learning, computer vision, language processing. The other major focus is on graphics processors both for general-purpose and graphics workloads.
The group consists of faculty members and students from Polytechnic University of Catalonia, University of Murcia and Rovira i Virgili University. The group has a long track record of scientific publications, with more than 500 research papers, and technology transfers, with more than 50 patents.
The research activities of the ARCO group focus on computer architecture, compilers and parallel processing, with special emphasis on microarchitecture and code generation techniques for energy-efficient and reliable computing systems. One of its main current focuses is on intelligent computing systems, where the goal is to devise novel architectures for machine learning, computer vision, language processing. The other major focus is on graphics processors both for general-purpose and graphics workloads.
The group consists of faculty members and students from Polytechnic University of Catalonia, University of Murcia and Rovira i Virgili University. The group has a long track record of scientific publications, with more than 500 research papers, and technology transfers, with more than 50 patents.
Collections in this community
-
Articles de revista [52]
-
Reports de recerca [11]
Recent Submissions
-
DRAM errors in the field: a statistical approach
(Association for Computing Machinery (ACM), 2019)
Conference report
Open AccessThis paper summarizes our two-year study of corrected and uncor-rected errors on the MareNostrum 3 supercomputer, covering 2000 billion MB-hours of DRAM in the field. The study analyzes 4.5 million corrected and 71 uncorrected ... -
Multi-objective interior design optimization method based on sustainability concepts for post-disaster temporary housing units
(2020-04)
Article
Restricted access - publisher's policyTemporary housing units (THUs), which are provided after disasters, are crucial in terms of sustainability pillars (economic, social, and environmental). In general, THUs, which are regular houses with minimum space and ... -
Neuron-level fuzzy memoization in RNNs
(Association for Computing Machinery (ACM), 2019)
Conference report
Open AccessRecurrent Neural Networks (RNNs) are a key technology for applications such as automatic speech recognition or machine translation. Unlike conventional feed-forward DNNs, RNNs remember past information to improve the ... -
Leveraging run-time feedback for efficient ASR acceleration
(Institute of Electrical and Electronics Engineers (IEEE), 2019)
Conference report
Open AccessIn this work, we propose Locality-AWare-Scheme (LAWS) for an Automatic Speech Recognition (ASR) accelerator in order to significantly reduce its energy consumption and memory requirements, by leveraging the locality among ... -
SCU: a GPU stream compaction unit for graph processing
(Association for Computing Machinery (ACM), 2019)
Conference report
Restricted access - publisher's policyGraph processing algorithms are key in many emerging applications in areas such as machine learning and data analytics. Although the processing of large scale graphs exhibits a high degree of parallelism, the memory access ... -
A low-power, high-performance speech recognition accelerator
(Institute of Electrical and Electronics Engineers (IEEE), 2019-12-01)
Article
Open AccessAutomatic Speech Recognition (ASR) is becoming increasingly ubiquitous, especially in the mobile segment. Fast and accurate ASR comes at high energy cost, not being affordable for the tiny power-budgeted mobile devices. ... -
CGPA: Coarse-Grained Pruning of Activations for Energy-Efficient RNN Inference
(2019-09-01)
Article
Open AccessRecurrent neural networks (RNNs) perform element-wise multiplications across the activations of gates. We show that a significant percentage of activations are saturated and propose coarse-grained pruning of activations ... -
Rendering elimination: early discard of redundant tiles in the graphics pipeline
(Institute of Electrical and Electronics Engineers (IEEE), 2019)
Conference report
Restricted access - publisher's policyGPUs are one of the most energy-consuming components for real-time rendering applications, since a large number of fragment shading computations and memory accesses are involved. Main memory bandwidth is especially taxing ... -
Early visibility resolution for removing ineffectual computations in the graphics pipeline
(Institute of Electrical and Electronics Engineers (IEEE), 2019)
Conference report
Restricted access - publisher's policyGPUs' main workload is real-time image rendering. These applications take a description of a (animated) scene and produce the corresponding image(s). An image is rendered by computing the colors of all its pixels. It is ... -
2018 International Symposium on Computer Architecture influential paper award
(2018-07-01)
Article
Open AccessThe International Symposium on Computer Architecture (ISCA) recognizes every year the most influential paper published in this conference 15 years earlier, based on its impact on research, development, products or ideas. ... -
Performance analysis and optimization of automatic speech recognition
(2018-10-01)
Article
Open AccessFast and accurate Automatic Speech Recognition (ASR) is emerging as a key application for mobile devices. Delivering ASR on such devices is challenging due to the compute-intensive nature of the problem and the power ... -
E-PUR: an energy-efficient processing unit for recurrent neural networks
(2018)
Conference report
Restricted access - publisher's policyRecurrent Neural Networks (RNNs) are a key technology for emerging applications such as automatic speech recognition, machine translation or image description. Long Short Term Memory (LSTM) networks are the most successful ...
RSS 1.0