Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 101
Encontrada(s) 2 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Exploring the interoperability of remote GPGPU virtualization using rCUDA and directive-based programming models

  • Castelló, Adrián|||0000-0002-8576-8451
  • Quintana-Ortí, Enrique S.|||0000-0002-5454-165X
  • Peña Monferrer, Antonio José
  • Mayo Gual, Rafael
  • Planas,Judit
  • Balaji, Pavan
[EN] Directive-based programming models, such as OpenMP, OpenACC, and OmpSs, enable users to accelerate applications by using coprocessors with little effort. These devices offer significant computing power, but their use can introduce two problems: an increase in the total cost of ownership and their underutilization because not all codes match their architecture. Remote accelerator virtualization frameworks address those problems. In particular, rCUDA provides transparent access to any graphic processor unit installed in a cluster, reducing the number of accelerators and increasing their utilization ratio. Joining these two technologies, directive-based programming models and rCUDA, is thus highly appealing. In this work, we study the integration of OmpSs and OpenACC with rCUDA, describing and analyzing several applications over three different hardware configurations that include two InfiniBand interconnections and three NVIDIA accelerators. Our evaluation reveals favorable performance results, showing low overhead and similar scaling factors when using remote accelerators instead of local devices., The researchers from the Universitat Jaume I de Castello were supported by Universitat Jaume I research project (P11B2013-21), project TIN2014-53495-R, a Generalitat Valenciana grant and FEDER. The researcher from the Barcelona Supercomputing Center (BSC-CNS) Lausanne was supported by the European Commission (HiPEAC-3 Network of Excellence, FP7-ICT 287759), Intel-BSC Exascale Lab collaboration, IBM/BSC Exascale Initiative collaboration agreement, Computacion de Altas Prestaciones VI (TIN2012-34557) and the Generalitat de Catalunya (2014-SGR-1051). This work was partially supported by the U.S. Dept. of Energy, Office of Science, Office of Advanced Scientific Computing Research (SC-21), under contract DE-AC02-06CH11357. The initial version of rCUDA was jointly developed by Universitat Politecnica de Valencia (UPV) and Universitat Jaume I de Castellon (UJI) until year 2010. This initial development was later split into two branches. Part of the UPV version was used in this paper. The development of the UPV branch was supported by Generalitat Valenciana under Grants PROMETEO 2008/060 and Prometeo II 2013/009. We gratefully acknowledge the computing resources provided and operated by the Joint Laboratory for System Evaluation (JLSE) at Argonne National Laboratory.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Two-sided orthogonal reductions to condensed forms on asymmetric multicore processors

  • Alonso-Jordá, Pedro|||0000-0002-6882-6592
  • Quintana-Ortí, Enrique S.|||0000-0002-5454-165X
  • Catalán, Sandra
  • Herrero, José R.
  • Rodríguez-Sánchez, Rafael
[EN] We investigate how to leverage the heterogeneous resources of an Asymmetric Multicore Processor (AMP) in order to deliver high performance in the reduction to condensed forms for the solution of dense eigenvalue and singular-value problems. The routines that realize this type of two-sided orthogonal reductions (TSOR) in LAPACK are especially challenging, since a significant fraction of their floating-point operations are cast in terms of memory-bound kernels while the remaining part corresponds to efficient compute-bound kernels. To deal with this scenario: (1) we leverage implementations of memory-bound and compute-bound kernels specifically tuned for AMPs; (2) we select the algorithmic block size for the TSOR routines via a practical model; and (3) we adjust the type and number of cores to use at each step of the reduction. Our experiments validate the model and assess the performance of our asymmetry-aware TSOR routines, using an ARMv7 big.LITTLE AMP, for three key operations: the reduction to tridiagonal form for symmetric eigenvalue problems, the reduction to Hessenberg form for non-symmetric eigenvalue problems, and the reduction to bidiagonal form for singular-value problems., The researchers from Universidad Jaume I were supported by project TIN2014-53495-R of MINECO and FEDER, and the FPU program of MECD. The researcher from Universitat Politecnica de Valencia was supported by the Generalitat Valenciana PROMETEOII/2014/003. The researcher from Universitat Politecnica de Catalunya was supported by projects TIN2015-65316-P from the Spanish Ministry of Education and 2014 SGR 1051 from the Generalitat de Catalunya, Dep. d'Innovacio, Universitats i Empresa.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Energy balance between voltage-frequency scaling and resilience for linear algebra routines on low-power multicore architectures

  • Catalán, Sandra
  • Herrero, José R.
  • Rodríguez-Sánchez, Rafael
  • Quintana-Ortí, Enrique S.|||0000-0002-5454-165X
[EN] Near Threshold Voltage (NTV) computing has been recently proposed as a technique to save energy, at the cost of incurring higher error rates including, among others, Silent Data Corruption (SDC). In this paper, we evaluate the energy efficiency of dense linear algebra routines using several low-power multicore processors and we analyze whether the potential energy reduction achieved when scaling the processor to operate at a low voltage compensates the cost of integrating a fault tolerance mechanism that tackles SDC. Our study targets algorithmic-based fault-tolerant versions of the dense matrix-vector and matrix(matrix) multiplication kernels (GEMV and GEMM, respectively), using the BLIS framework, as well as an implementation of the LU factorization with partial pivoting built on top of GEMM, Furthermore, we tailor the study for a number of representative 32-bit and 64-bit multicore processors from ARM that were specifically designed for energy efficiency. (C) 2017 Elsevier B.V. All rights reserved., The researchers from Universidad Jaume I were supported by project CICYT TIN2014-53495-R of MINECO and FEDER, and the FPU program of MECD. The researcher from Universitat Politecnica de Catalunya was supported by projects TIN2015-65316-P from the Spanish Ministry of Education and 2014 SGR 1051 from the Generalitat de Catalunya, Dep. d'Innovacio, Universitats i Empresa.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Multi-threaded dense linear algebra libraries for low-power asymmetric multicore processors

  • Catalán, Sandra
  • Herrero, José R.
  • Igual Peña, Francisco Daniel
  • Rodríguez-Sánchez, Rafael
  • Adeniyi-Jones, Chris
  • Quintana-Ortí, Enrique S.|||0000-0002-5454-165X
[EN] Dense linear algebra libraries, such as BLAS and LAPACK, provide a relevant collection of numerical tools for many scientific and engineering applications. While there exist high performance implementations of the BLAS (and LAPACK) functionality for many current multi-threaded architectures, the adaption of these libraries for asymmetric multicore processors (AMPs) is still pending. In this paper we address this challenge by developing an asymmetry-aware implementation of the BLAS, based on the BLIS framework, and tailored for AMPs equipped with two types of cores: fast/power-hungry versus slow/energy-efficient. For this purpose, we integrate coarse-grain and fine-grain parallelization strategies into the library routines which, respectively, dynamically distribute the workload between the two core types and statically repartition this work among the cores of the same type. Our results on an ARM (R) big.LITTLE (TM) processor embedded in the Exynos 5422 SoC, using the asymmetry-aware version of the BLAS and a plain migration of the legacy version of LAPACK, experimentally assess the benefits, limitations, and potential of this approach from the perspectives of both throughput and energy efficiency. (C) 2016 Elsevier B.V. All rights reserved., The researchers from Universidad Jaume I were supported by projects CICYT TIN2011-23283 and TIN2014-53495-R of MINECO and FEDER, and the FPU program of MECD. The researcher from Universidad Complutense de Madrid was supported by project CICYT TIN2015-65277-R. The researcher from Universitat Politecnica de Catalunya was supported by projects TIN2015-65316-P from the Spanish Ministry of Education and 2014 SGR 1051 from the Generalitat de Catalunya, Dep. dinnovacio, Universitats i Empresa.
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Low-latency multi-threaded ensemble learning for dynamic big data streams

  • Marron, Diego
  • Ayguadé Parra, Eduard
  • Herrero Zaragoza, José Ramón
  • Read, Jesse
  • Bifet Figuerol, Albert Carles
© 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Real–time mining of evolving data streams involves new challenges when targeting today’s application domains such as the Internet of the Things: increasing volume, velocity and volatility requires data to be processed on–the–fly with fast reaction and adaptation to changes. This paper presents a high performance scalable design for decision trees and ensemble combinations that makes use of the vector SIMD and multicore capabilities available in modern processors to provide the required throughput and accuracy. The proposed design offers very low latency and good scalability with the number of cores on commodity hardware when compared to other state–of–the art implementations. On an Intel i7-based system, processing a single decision tree is 6x faster than MOA (Java), and 7x faster than StreamDM (C++), two well- known reference implementations. On the same system, the use of the 6 cores (and 12 hardware threads) available allow to process an ensemble of 100 learners 85x faster that MOA while providing the same accuracy. Furthermore, our solution is highly scalable: on an Intel Xeon socket with large core counts, the proposed ensemble design achieves up to 16x speed-up when employing 24 cores with respect to a single threaded execution., This work is partially supported by the Spanish Government through Programa Severo Ochoa (SEV-2015-0493), by the Spanish Ministry of Science and Technology through TIN2015-65316-P project, by the Generalitat de Catalunya (contract 2014-SGR-1051), by the Universitat Politècnica de Catalunya through an FPI/UPC scholarship and by NVIDIA through the UPC/BSC GPU Center of Excellence., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Extending the OpenCHK Model with advanced checkpoint features

  • Maroñas Bravo, Marcos
  • Mateo Bellido, Sergi
  • Keller, Kai Rasmus
  • Bautista Gomez, Leonardo
  • Ayguadé Parra, Eduard
  • Beltran Querol, Vicenç
One of the major challenges in using extreme scale systems efficiently is to mitigate the impact of faults. Application-level checkpoint/restart (CR) methods provide the best trade-off between productivity, robustness, and performance. There are many solutions implementing CR at the application level. They all provide advanced I/O capabilities to minimize the overhead introduced by CR. Nevertheless, there is still room for improvement in terms of programmability and flexibility, because end-users must manually serialize and deserialize application state using low-level APIs, modify the flow of the application to consider restarts, or rewrite CR code whenever the backend library changes. In this work, we propose a set of compiler directives and clauses that allow users to specify CR operations in a simple way. Our approach supports the common CR features provided by all the CR libraries. However, it can also be extended to support advanced features that are only available in some CR libraries, such as differential checkpointing, the use of HDF5 format, and the possibility of using fault-tolerance-dedicated threads. The result of our evaluation revealed a high increase in programmability. On average, we reduced the number of lines of code by 71%, 94%, and 64% for FTI, SCR, and VeloC, respectively, and no additional overhead was perceived using our solution compared to using the backend libraries directly. Finally, portability is enhanced because our programming model allows the use of any backend library without changing any code., This work is supported by the Spanish Ministerio de Ciencia, Innovación y Universidades (TIN2015-65316-P) and by the Generalitat de Catalunya (2014-SGR-1051). This project received funding from the European Union’s Seventh Framework Programme (FP7/2007-2013) and the Horizon 2020 (H2020) funding framework under grant agreement no. H2020-FETHPC-754304 (DEEPEST)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Look-ahead in the two-sided reduction to compact band forms for symmetric eigenvalue problems and the SVD

  • Rodríguez Sánchez, Rafael
  • Catalán Pallarés, Sandra
  • Herrero Zaragoza, José Ramón
  • Quintana Ortí, Enrique Salvador
  • Tomás Domínguez, Andrés Enrique
We address the reduction to compact band forms, via unitary similaritytransformations, for the solution of symmetric eigenvalue problems and the compu-tation of the singular value decomposition (SVD). Concretely, in the first case, werevisit the reduction to symmetric band form, while, for the second case, we pro-pose a similar alternative, which transforms the original matrix to (unsymmetric)band form, replacing the conventional reduction method that produces a triangular–band output. In both cases, we describe algorithmic variants of the standard Level3 Basic Linear Algebra Subroutines (BLAS)-based procedures, enhanced with look-ahead, to overcome the performance bottleneck imposed by the panel factorization., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Work-efficient parallel non-maximum suppression kernels

  • Oro García, David
  • Fernandez Tena, Carles
  • Martorell Bofill, Xavier
  • Hernando Pericás, Francisco Javier
In the context of object detection, sliding-window classifiers and single-shot convolutional neural network (CNN) meta-architectures typically yield multiple overlapping candidate windows with similar high scores around the true location of a particular object. Non-maximum suppression (NMS) is the process of selecting a single representative candidate within this cluster of detections, so as to obtain a unique detection per object appearing on a given picture. In this paper, we present a highly scalable NMS algorithm for embedded graphics processing unit (GPU) architectures that is designed from scratch to handle workloads featuring thousands of simultaneous detections on a given picture. Our kernels are directly applicable to other sequential NMS algorithms such as FeatureNMS, Soft-NMS or AdaptiveNMS that share the inner workings of the classic greedy NMS method. The obtained performance results show that our parallel NMS algorithm is capable of clustering 1024 simultaneous detected objects per frame in roughly 1 ms on both Tegra X1 and Tegra X2 on-die GPUs, while taking 2 ms on Tegra K1. Furthermore, our proposed parallel greedy NMS algorithm yields a 14–40x speed up when compared to state-of-the-art NMS methods that require learning a CNN from annotated data., This work has been partially supported by the Ministerio de Economía y Competitividad under contracts (TIN2015-65316-P, TEC2012-38939-C03-02), the Departament d’Innovació, Universitats i Empresa de la Generalitat de Catalunya under project MPEXPAR: Models de Programació i Entorns d’Execució Paral·lels (2014-SGR-1051), and the European Commission under the Horizon 2020 program (H2020-ICT-644312)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Runtime-guided management of stacked DRAM memories in task parallel programs

  • Álvarez Martí, Lluc
  • Casas, Marc
  • Labarta Mancho, Jesús José
  • Ayguadé Parra, Eduard
  • Valero Cortés, Mateo
  • Moretó Planas, Miquel
Stacked DRAM memories have become a reality in High-Performance Computing (HPC) architectures. These memories provide much higher bandwidth while consuming less power than traditional off-chip memories, but their limited memory capacity is insufficient for modern HPC systems. For this reason, both stacked DRAM and off-chip memories are expected to co-exist in HPC architectures, giving raise to different approaches for architecting the stacked DRAM in the system. This paper proposes a runtime approach to transparently manage stacked DRAM memories in task-based programming models. In this approach the runtime system is in charge of copying the data accessed by the tasks to the stacked DRAM, without any complex hardware support nor modifications to the application code. To mitigate the cost of copying data between the stacked DRAM and the off-chip memory, the proposal includes an optimization to parallelize the copies across idle or additional helper threads. In addition, the runtime system is aware of the reuse pattern of the data accessed by the tasks, and can exploit this information to avoid unworthy copies of data to the stacked DRAM. Results on the Intel Knights Landing processor show that the proposed techniques achieve an average speedup of 14% against the state-of-the-art library to manage the stacked DRAM and 29% against a stacked DRAM architected as a hardware cache., This work has been supported by the RoMoL ERC Advanced Grant (GA 321253), by the European HiPEAC Network of Excellence, by the Spanish Ministry of Economy and Competitiveness (contract TIN2015-65316-P), by the Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272) and by the European Union’s Horizon 2020 research and innovation programme (grant agreement 779877). M. Moreto has been partially supported by the Spanish Ministry of Economy, Industry and Competitiveness under Ramon y Cajal fellowship number RYC-2016-21104., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Reducing data movement on large shared memory systems by exploiting computation dependencies

  • Barrera, I.S.
  • Ayguadé Parra, Eduard
  • Valero Cortés, Mateo
  • Moretó Planas, Miquel
  • Labarta Mancho, Jesús José
  • Casas, Marc
Shared memory systems are becoming increasingly complex as they typically integrate several storage devices. That brings different access latencies or bandwidth rates depending on the proximity between the cores where memory accesses are issued and the storage devices containing the requested data. In this context, techniques to manage and mitigate non-uniform memory access (NUMA) effects consist in migrating threads, memory pages or both and are generally applied by the system software. We propose techniques at the runtime system level to further mitigate the impact of NUMA effects on parallel applications' performance. We leverage runtime system metadata expressed in terms of a task dependency graph, where nodes are pieces of serial code and edges are control or data dependencies between them, to efficiently reduce data transfers. Our approach, based on graph partitioning, adds negligible overhead and is able to provide performance improvements up to 1.52× and average improvements of 1.12× with respect to the best state-of-the-art approach when deployed on a 288-core shared-memory system. Our approach reduces the coherence traffic by 2.28× on average with respect to the state-of-the-art., This work has been supported by the RoMoL ERC Advanced Grant (GA 321253), by the European HiPEAC Network of Excellence, by the Spanish Ministry of Economy and Competitiveness (contract TIN2015-65316-P), by the Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272) and by the European Union’s Horizon 2020 research and innovation programme (grant agreements 671697 and 779877). I. Sánchez Barrera has been partially supported by the Spanish Ministry of Education, Culture and Sport under Formación del Profesorado Universitario fellowship number FPU15/03612. M. Moretó has been partially supported by the Spanish Ministry of Economy, Industry and Competitiveness under Ramón y Cajal fellowship number RYC-2016-21104., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Towards an auto-tuned and task-based SpMV (LASs Library)

  • Catalán Pallarés, Sandra
  • Usui, Tetsuzo
  • Toledo, Leonel
  • Martorell Bofill, Xavier
  • Labarta Mancho, Jesús José
  • Valero Lara, Pedro
We present a novel approach to parallelize the SpMV kernel included in LASs (Linear Algebra routines on OmpSs) library, after a deep review and analysis of several well-known approaches. LASs is based on OmpSs, a task-based runtime that extends OpenMP directives, providing more flexibility to apply new strategies. Based on tasking and nesting, with the aim of improving the workload imbalance inherent to the SpMV operation, we present a strategy especially useful for highly imbalanced input matrices. In this approach, the number of created tasks is dynamically decided in order to maximize the use of the resources of the platform. Throughout this paper, SpMV behavior depending on the selected strategy (state of the art and proposed strategies) is deeply analyzed, setting in this way the base for a future auto-tunable code that is able to select the most suitable approach depending on the input matrix. The experiments of this work were carried out for a set of 12 matrices from the Suite Sparse Matrix Collection, all of them with different characteristics regarding their sparsity. The experiments of this work were performed on a node of Marenostrum 4 supercomputer (with two sockets Intel Xeon, 24 cores each) and on a node of Dibona cluster (using one ARM ThunderX2 socket with 32 cores). Our tests show that, for Intel Xeon, the best parallelization strategy reduces the execution time of the reference MKL multi-threaded version up to 67%. On ARM ThunderX2, the reduction is up to 56% with respect to the OmpSs parallel reference., This project has received funding from the Spanish Ministry of Economy and Competitiveness under the project Computación de Altas Prestaciones VII (TIN2015- 65316-P), the Departament d’Innovació, Universitats i Empresa de la Generalitat de Catalunya, under project MPEXPAR: Models de Programació i Entorns d’Execució Parallels (2014-SGR-1051), and the Juan de la Cierva Grant Agreement No IJCI-2017- 33511, and the Spanish Ministry of Science and Innovation under the project Heterogeneidad y especialización en la era post-Moore (RTI2018-093684-BI00). We also acknowledge the funding provided by Fujitsu under the BSC-Fujitsu joint project: Math Libraries Migration and Optimization, Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Performance characterization of video analytics workloads in heterogeneous edge infrastructures

  • Rivas Barragan, Daniel
  • Guim Bernat, Francesc
  • Polo Bardés, Jorda
  • Carrera Pérez, David
Powered by deep learning, video analytic applications process millions of camera feeds in real-time to extract meaningful information from their surroundings. And this number grows by the minute. To avoid saturating the backhaul network and provide lower latencies, a distributed and heterogeneous edge cloud is postulated as a key enabler for widespread video analytics. This article provides a complete characterization of end-to-end video analytics across a set of hardware platforms and different neural network architectures. Each platform is selected to fill a different gap in a distributed, shared, and heterogeneous infrastructure. Moreover, we analyze how performance scales on each of these platforms with respect to the amount of resources dedicated to video analytics. Finally, we extract the key conclusions of the characterization to build an experimental model to estimate performance and cost of end-to-end video analytics in different edge scenarios. Our experiments show that managing video analytics workloads efficiently requires awareness of both, the platforms in which these are executed, and the full end-to-end pipeline. To the best of our knowledge, this is the first work that provides a complete characterization of end-to-end video analytics in heterogeneous edge platforms., This work was supported by the Ministry of Economy of Spain under contract TIN2015-65316-P and the Ministry of Science under contract PID2019-107255GB-C21/AEI/10.13039/501100011033; the Generalitat de Catalunya under contract 2014SGR1051; the ICREA Academia program; and Intel Corporation, Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Adaptive request scheduling for the I/O forwarding layer using reinforcement learning

  • Bez, Jean Luca
  • Zanon Boito, Francieli
  • Nou Castell, Ramon
  • Miranda Bueno, Alberto
  • Cortés, Toni
  • Navaux, Philippe O.A.
In this paper, we propose an approach to adapt the I/O forwarding layer of HPC systems to applications’ access patterns. I/O optimization techniques can improve performance for the access patterns they were designed to target, but they often decrease performance for others. Furthermore, these techniques usually depend on the precise tune of their parameters, which commonly falls back to the users. Instead, we propose to do it dynamically at runtime based on the I/O workload observed by the system. Our approach uses a reinforcement learning technique – contextual bandits – to make the system capable of learning the best parameter value to each observed access pattern during its execution. That eliminates the need of a complicated and time-consuming previous training phase. Our case study is the TWINS scheduling algorithm, where performance improvements depend on the time window parameter, which in turn depends on the workload. We evaluate our proposal and demonstrate it can reach a precision of 88% on the parameter selection in the first hundreds of observations of an access pattern, achieving 99% of the optimal performance. We demonstrate that the system – which is expected to live for years – will be able to adapt to changes and optimize its performance after having observed an access pattern for a few (not necessarily contiguous) minutes., This study was financed in part by the Coordenaçao de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. It has also received support from the Conselho Nacional de Desenvolvimento Científico e Tecnologico (CNPq), Brazil; the LICIA International Laboratory; NCSA-Inria-ANL-BSC-JSC-Riken Joint-Laboratory on Extreme Scale Computing (JLESC); the Spanish Ministry of Science and Innovation under the TIN2015–65316 grant; and the Generalitat de Catalunya under contract 2014– SGR–1051. This research received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 800144., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Adaptive prediction models for data center resources utilization estimation

  • Baig, Shuja-ur-Rehman
  • Iqbal, Waheed
  • Berral García, Josep Lluís
  • Erradi, Abdelkarim
  • Carrera Pérez, David
Accurate estimation of data center resource utilization is a challenging task due to multi-tenant co-hosted applications having dynamic and time-varying workloads. Accurate estimation of future resources utilization helps in better job scheduling, workload placement, capacity planning, proactive auto-scaling, and load balancing. The inaccurate estimation leads to either under or over-provisioning of data center resources. Most existing estimation methods are based on a single model that often does not appropriately estimate different workload scenarios. To address these problems, we propose a novel method to adaptively and automatically identify the most appropriate model to accurately estimate data center resources utilization. The proposed approach trains a classifier based on statistical features of historical resources usage to decide the appropriate prediction model to use for given resource utilization observations collected during a specific time interval. We evaluated our approach on real datasets and compared the results with multiple baseline methods. The experimental evaluation shows that the proposed approach outperforms the state-of-the-art approaches and delivers 6% to 27% improved resource utilization estimation accuracy compared to baseline methods., This work is partially supported by the European Research Council (ERC) under the EU Horizon 2020 programme (GA 639595), the Spanish Ministry of Economy, Industry and Competitiveness (TIN2015-65316-P and IJCI2016-27485), the Generalitat de Catalunya (2014-SGR-1051), and NPRP grant # NPRP9-224-1-049 from the Qatar National Research Fund (a member of Qatar Foundation) and University of the Punjab, Pakistan., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Task Packing: Efficient task scheduling in unbalanced parallel programs to maximize CPU utilization

  • Utrera Iglesias, Gladys Miriam
  • Farreras Esclusa, Montse
  • Fornés de Juan, Jordi
Load imbalance in parallel systems can be generated by external factors to the currently running applications like operating system noise or the underlying hardware like a heterogeneous cluster. HPC applications working on irregular data structures can also have difficulties to balance their computations across the parallel tasks. In this article we extend, improve and evaluate more deeply the Task Packing mechanism proposed in a previous work. The main idea of the mechanism is to concentrate the idle cycles of unbalanced applications in such a way that one or more CPUs are freed from execution. To achieve this, CPUs are stressed with just useful work of the parallel application tasks, provided performance is not degraded. The packing is solved by an algorithm based on the Knapsack problem, in a minimum number of CPUs and using oversubscription. We design and implement a more efficient version of such mechanism. To that end, we perform the Task Packing “in place”, taking advantage of idle cycles generated at synchronization points of unbalanced applications. Evaluations are carried out on a heterogeneous platform using FT and miniFE benchmarks. Results showed that our proposal generates low overhead. In addition the amount of freed CPUs are related to a load imbalance metric which can be used as a prediction for it., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Intelligent adaptation of hardware knobs for improving performance and power consumption

  • Ortega Carrasco, Cristobal
  • Álvarez Martí, Lluc
  • Casas, Marc
  • Bertran, Ramon
  • Buyuktosunoglu, Alper
  • Eichenberger, Alexandre
  • Bose, Pradip
  • Moretó Planas, Miquel
Current microprocessors include several knobs to modify the hardware behavior in order to improve performance, power, and energy under different workload demands. An impractical and time consuming offline profiling is needed to evaluate the design space to find the optimal knob configuration. Different knobs are typically configured in a decoupled manner to avoid the time-consuming offline profiling process. This can often lead to underperforming configurations and conflicting decisions that jeopardize system power-performance efficiency. Thus, a dynamic management of the different hardware knobs is necessary to find the knob configuration that maximizes system power-performance efficiency without the burden of offline profiling. In this paper, we propose libPRISM, an infrastructure that enables the transparent management of multiple hardware knobs in order to adapt the system to the evolving demands of hardware resources in different workloads. libPRISM can minimize execution time, energy-delay product or power consumption by dynamically managing the SMT level, the data prefetcher, and the DVFS hardware knobs. Overall, the proposed solutions increase performance up to 130% (16.9% on average), reduce energy-delay product up to 80%, and reduce power consumption up to 33% depending on the target metric compared to the default knob configuration of the system., This work has been supported by the RoMoL ERC Advanced Grant (GA 321253), by the European HiPEAC Network of Excellence, by the Spanish Ministry of Science and Innovation (contracts TIN2015-65316-P), by Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272) and by IBM/BSC Deep Learning Center initiative.This research was developed in part with funding from the Defense Advanced Research Projects Agency (DARPA)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

The AXIOM project: IoT on heterogeneous embedded platforms

  • Filgueras Izquierdo, Antonio
  • Vidal, Miquel
  • Mateu, Marc
  • Jiménez González, Daniel
  • Álvarez Martínez, Carlos
  • Martorell Bofill, Xavier
  • Ayguadé Parra, Eduard
  • Theodoropoulos, Dimitris
  • Pnevmatikatos, Dionisis
  • Gai, Paolo
  • Garzarella, Stefano
  • Oro de Herta, David
  • Hernando Pericás, Francisco Javier
  • Bettin, Nicola
  • Pomella, Alberto
  • Giorgi, Roberto
© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., The AXIOM project aims at providing an environment for Cyber-Physical Systems. Smart Video Surveillance targets public environments, involving real-time face detection in crowds. Smart Home Living targets home environments and access control. These applications are used as experimental usecases for the AXIOM platform, currently based on the Xilinx Zynq-7000 SoCs. We have integrated the Xilinx Vivado HLS tool for the FPGA support within the OmpSs programming model, to enable OpenMP-like programming in the FPGA. This paper presents the programming environment, and the evaluation of the most computationally expensive parts of the target applications., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

GekkoFS: A temporary burst buffer file system for HPC applications

  • Vef, Marc-André
  • Moti, Nafiseh
  • Süb, Tim
  • Tacke, Markus
  • Tocci, Tommaso
  • Nou, Ramon
  • Miranda Bueno, Alberto
  • Cortés, Toni
  • Brinkmann, Andre
Many scientific fields increasingly use high-performance computing (HPC) to process and analyze massive amounts of experimental data while storage systems in today’s HPC environments have to cope with new access patterns. These patterns include many metadata operations, small I/O requests, or randomized file I/O, while general-purpose parallel file systems have been optimized for sequential shared access to large files. Burst buffer file systems create a separate file system that applications can use to store temporary data. They aggregate node-local storage available within the compute nodes or use dedicated SSD clusters and offer a peak bandwidth higher than that of the backend parallel file system without interfering with it. However, burst buffer file systems typically offer many features that a scientific application, running in isolation for a limited amount of time, does not require. We present GekkoFS, a temporary, highly-scalable file system which has been specifically optimized for the aforementioned use cases. GekkoFS provides relaxed POSIX semantics which only offers features which are actually required by most (not all) applications. GekkoFS is, therefore, able to provide scalable I/O performance and reaches millions of metadata operations already for a small number of nodes, significantly outperforming the capabilities of common parallel file systems., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Two-sided orthogonal reductions to condensed forms on asymmetric multicore processors

  • Alonso Jordá, Pedro
  • Catalán Pallarés, Sandra
  • Herrero Zaragoza, José Ramón
  • Quintana Ortí, Enrique Salvador
  • Rodríguez Sánchez, Rafael
We investigate how to leverage the heterogeneous resources of an Asymmetric Multicore Processor (AMP) in order to deliver high performance in the reduction to condensed forms for the solution of dense eigenvalue and singular-value problems. The routines that realize this type of two-sided orthogonal reductions (TSOR) in LAPACK are especially challenging, since a significant fraction of their floating-point operations are cast in terms of memory-bound kernels while the remaining part corresponds to efficient compute-bound kernels. To deal with this scenario: (1) we leverage implementations of memory-bound and compute-bound kernels specifically tuned for AMPs; (2) we select the algorithmic block size for the TSOR routines via a practical model; and (3) we adjust the type and number of cores to use at each step of the reduction. Our experiments validate the model and assess the performance of our asymmetry-aware TSOR routines, using an ARMv7 big.LITTLE AMP, for three key operations: the reduction to tridiagonal form for symmetric eigenvalue problems, the reduction to Hessenberg form for non-symmetric eigenvalue problems, and the reduction to bidiagonal form for singular-value problems., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Automatic distributed deep learning using resource-constrained edge devices

  • Gutiérrez Torre, Alberto
  • Bahadori, Kiyana
  • Baig, Shuja-ur-Rehman
  • Iqbal, Waheed
  • Vardanega, Tullio
  • Berral García, Josep Lluís
  • Carrera Pérez, David
Processing data generated at high volume and speed from the Internet of Things, smart cities, domotic, intelligent surveillance, and e-healthcare systems require efficient data processing and analytics services at the Edge to reduce the latency and response time of the applications. The Fog Computing Edge infrastructure consists of devices with limited computing, memory, and bandwidth resources, which challenge the construction of predictive analytics solutions that require resource-intensive tasks for training machine learning models. In this work, we focus on the development of predictive analytics for urban traffic. Our solution is based on deep learning techniques localized in the Edge, where computing devices have very limited computational resources. We present an innovative method for efficiently training of Gated Recurrent-Units (GRUs) across available resource-constrained CPU and GPU Edge devices. Our solution employs distributed GRU model learning and dynamically stops the training process to utilize the low-power and resource-constrained Edge devices while ensuring good estimation accuracy effectively. The proposed solution was extensively evaluated using low-powered ARM-based devices, including Raspberry Pi v3 and the low-powered GPU-enabled device NVIDIA Jetson Nano, and also compared them with Single-CPU Intel Xeon machines. For the evaluation experiments, we used real-world Floating Car Data. The experiments show that the proposed solution delivers excellent prediction accuracy and computational performance on the Edge when compared with the baseline methods., This work has been partially supported by the Spanish Government (contract PID2019-107255GB), the Generalitat de Catalunya (contract 2014-SGR-1051), the University of Padova, and Severo Ochoa CoE (SEV-2015-0493-16-5). The statements made herein are solely the responsibility of the authors., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Long short-term memory networks for earthquake detection in Venezuelan regions

  • Mus León, Sergi
  • Gutiérrez Escobar, Norma
  • Tous Liesa, Rubén
  • Otero Calviño, Beatriz
  • Cruz de la Cruz, Stalin Leonel
  • Llácer Giner, David
  • Alvarado Bermúdez, Leonardo
  • Rojas, Otilio
Reliable earthquake detection and location algorithms are necessary to properly catalog and analyze the continuously growing seismic records. This paper reports the results of applying Long Short-Term Memory (LSTM) networks to single-station three-channel waveforms for P-wave earthquake detection in western and north central regions of Venezuela. Precisely, we apply our technique to study the seismicity along the dextral strike-slip Boconó and La Victoria - San Sebastián faults, with complex tectonics driven by the interactions between the South American and Caribbean plates., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

MPI+OpenMP tasking scalability for multi-morphology simulations of the human brain

  • Valero-Lara, Pedro
  • Sirvent, Raül
  • Peña, Antonio J.
  • Labarta Mancho, Jesús José
The simulation of the behavior of the human brain is one of the most ambitious challenges today with a non-end of important applications. We can find many different initiatives in the USA, Europe and Japan which attempt to achieve such a challenging target. In this work, we focus on the most important European initiative (the Human Brain Project) and on one of the models developed in this project. This tool simulates the spikes triggered in a neural network by computing the voltage capacitance on the neurons’ morphology, being one of the most precise simulators today. In the present work, we have evaluated the use of MPI+OpenMP tasking on top of this framework. We prove that this approach is able to achieve a good scaling even when computing a relatively low workload (number of neurons) per node. One of our targets consists of achieving not only a highly scalable implementation, but also to develop a tool with a high degree of abstraction without losing control and performance by using MPI+OpenMP tasking. The main motivation of this work is the evaluation of this cutting-edge simulation on multi-morphology neural networks. The simulation of a high number of neurons, which are completely different among them, is an important challenge. In fact, in the multi-morphology simulations, we find an important unbalancing between the nodes, mainly due to the differences in the neurons, which causes an important under-utilization of the available resources. In this work, the authors present and evaluate mechanisms to deal with this and reduce the time of this kind of simulations considerably., We would like to appreciate the valuable feedback and help provided by the main developers of the Arbor simulator: Benjamin Cumming (ETH Zürich) and Alexander Peyser (Jülich Supercomputing Center). This project has also received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 720270 (HBP SGA1 and SGA2), from the Spanish Ministry of Economy and Competitiveness under the project Computación de Altas Prestaciones VII (TIN2015-65316- P) and the Departament d’Innovació, Universitats i Empresa de la Generalitat de Catalunya, under project MPEXPAR: Models de Programació i Entorns d’Execució Paral · lels (2014-SGR-1051). Finally, this project also received funding from the Spanish Ministry of Economy and Competitiveness under the Juan de la Cierva Grant Agreement No IJCI-2017-33511, and from the European Union’s Horizon 2020 research and innovation program under the Marie Sklodowska Curie grant agreement No. 749516., Peer Reviewed, Postprint (author's final draft), Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

The impact of non-additive genetic associations on age-related complex diseases

  • Guindo Martínez, Marta
  • Amela Milian, Ramon
  • Bonàs Guarch, Silvia
  • Puiggros Maldonado, Montserrat
  • Salvoro, Cecilia
  • Miguel Escalada, Irene
  • Sánchez Castaño, Friman
  • Ramón Cortés, Cristian
  • Badia Sala, Rosa Maria
  • Ejarque Artigas, Jorge
  • Mercader Bigas, Josep Maria
  • Torrents Rodas, David
Genome-wide association studies (GWAS) are not fully comprehensive, as current strategies typically test only the additive model, exclude the X chromosome, and use only one reference panel for genotype imputation. We implement an extensive GWAS strategy, GUIDANCE, which improves genotype imputation by using multiple reference panels and includes the analysis of the X chromosome and non-additive models to test for association. We apply this methodology to 62,281 subjects across 22 age-related diseases and identify 94 genome-wide associated loci, including 26 previously unreported. Moreover, we observe that 27.7% of the 94 loci are missed if we use standard imputation strategies with a single reference panel, such as HRC, and only test the additive model. Among the new findings, we identify three novel low-frequency recessive variants with odds ratios larger than 4, which need at least a three-fold larger sample size to be detected under the additive model. This study highlights the benefits of applying innovative strategies to better uncover the genetic architecture of complex diseases., This work has been sponsored by the grant SEV-2011-00067 and SEV2015-0493 of Severo Ochoa Program, awarded by the Spanish Government, by the grant TIN2015- 65316-P, awarded by the Spanish Ministry of Science and Innovation, and by the Generalitat de Catalunya (contract 2014-SGR-1051). This work was supported by an EFSD/Lilly research fellowship. Josep M. Mercader was supported by a Sara Borrell Fellowship from the Instituto Carlos III, Beatriu de Pinós fellowship from the Agency for Management of University and Research Grants (AGAUR) and by the American Diabetes Association Innovative and Clinical Translational Award 1-19-ICTS-068. Sílvia Bonàs was supported by FI-DGR Fellowship from FIDGR 2013 from Agència de Gestió d’Ajuts Universitaris i de Recerca (AGAUR, Generalitat de Catalunya), and a ‘Juan de la Cierva’ postdoctoral fellowship (MINECO;FJCI-2017-32090). Cecilia Salvoro received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement H2020-MSCA-COFUND-2016- 754433. Cristian Ramon-Cortes pre-doctoral contract is financed by the Spanish Ministry of Science, Innovation, and Universities under contract BES-2016-076791. Elizabeth G. Atkinson was supported by the National Institutes of Mental Health (grants K01MH121659 and T32MH017119). Jose Florez was supported by NIH/NIDDK award K24 DK110550. This study made use of data generated by the UK10K Consortium, derived from samples from UK10K COHORT IMPUTATION (EGAS00001000713). A full list of the investigators who contributed to the generation of the data is available at www.UK10K.org. Funding for UK10K was provided by the Wellcome Trust under award WT091310. This study made use of data generated by the ‘Genome of the Netherlands’ project, which is funded by the Netherlands Organization for Scientific Research (grant no. 184021007). The data were made available as a Rainbow Project of BBMRI-NL. Samples were contributed by LifeLines (http://lifelines.nl/lifelines-research/general), the Leiden Longevity Study (http://www.healthy-ageing.nl; http://www.langleven.net), the Netherlands Twin Registry (NTR: http://www.tweelingenregister.org), the Rotterdam studies (http://www.erasmus-epidemiology.nl/rotterdamstudy) and the Genetic Research in Isolated Populations program (http://www.epib.nl/research/geneticepi/research. html#gip). The sequencing was carried out in collaboration with the Beijing Institute for Genomics (BGI). This study also made use of data generated by The Haplotype Reference Consortium (HRC) accessed through The European Genome-phenome Archive at the European Bioinformatics Institute with the accession numbers EGAD00001002729, after a form agreed by the Barcelona Supercomputing Center (BSC) with WTSI. This research has been conducted using also the UK Biobank Resource (application number 31063 and 27892). The Genotype-Tissue Expression (GTEx) Project was supported by the Common Fund of the Office of the Director of the National Institutes of Health, and by NCI, NHGRI, NHLBI, NIDA, NIMH, and NINDS. The data used for the analyses described in this manuscript were obtained from the GTEx Portal on 07/16/2019. We acknowledge PRACE for awarding us access to both MareNostrum supercomputer from the Barcelona Supercomputing Center, based in Spain at Barcelona, and the SuperMUC supercomputer of the Leibniz Supercomputing Center (LRZ), based in Garching at Germany (proposals numbers 2016143358 and 2016163985). The technical support group from the Barcelona Supercomputing Center is gratefully acknowledged. Finally, we thank all the Computational Genomics group at the BSC for their helpful discussions and valuable comments on the manuscript. We also acknowledge Elias Rodriguez Fos for designing the GUIDANCE logo., Peer Reviewed, Article signat per 22 autors/autores: Marta Guindo-Martínez 1,18; Ramon Amela 1,18; Silvia Bonàs-Guarch 1,2,3; Montserrat Puiggròs 1; Cecilia Salvoro 1; Irene Miguel-Escalada 1,2,3; Caitlin E. Carey 4,5; Joanne B. Cole 6,7,8,9; Sina Rüeger 10; Elizabeth Atkinson 4,5,11; Aaron Leong 8,12; Friman Sanchez 1; Cristian Ramon-Cortes 1; Jorge Ejarque 1; Duncan S. Palmer 4,5,17; Mitja Kurki 10; FinnGen Consortium*, Krishna Aragam 11,13,14; Jose C. Florez 6,7,15; Rosa M. Badia 1; Josep M. Mercader 1,6,7,15,19✉ & David Torrents 1,16,19✉ *A full list of members and their affiliations appears in the Supplementary Information 1 Barcelona Supercomputing Center (BSC), Barcelona, Spain. 2 Regulatory Genomics and Diabetes, Centre for Genomic Regulation, The Barcelona Institute of Science and Technology, Barcelona, Spain. 3 CIBER de Diabetes y Enfermedades Metabólicas Asociadas, Madrid, Spain. 4 Stanley Center for Psychiatric Research, Broad Institute of MIT and Harvard, Cambridge, MA, USA. 5 Analytic and Translational Genetics Unit, Department of Medicine, Massachusetts General Hospital, Boston, MA, USA. 6 Programs in Metabolism and Medical and Population Genetics, Broad Institute of MIT and Harvard, Cambridge, MA, USA. 7 Diabetes Unit and Center for Genomic Medicine, Massachusetts General Hospital, Boston, MA, USA. 8 Harvard Medical School, Boston, MA, USA. 9 Division of Endocrinology and Center for Basic and Translational Obesity Research, Boston Children’s Hospital, Boston, MA, USA. 10 Institute for Molecular Medicine Finland, FIMM, HiLIFE, University of Helsinki, Helsinki, Finland. 11 Program in Medical and Population Genetics, Broad Institute of MIT and Harvard, Cambridge, MA, USA. 12 Department of Medicine, Massachusetts General Hospital, Boston, MA, USA. 13 Cardiology Division, Massachusetts General Hospital, Boston, MA, USA. 14 Cardiovascular Research Center, Massachusetts General Hospital, Boston, MA, USA. 15 Department of Medicine, Harvard Medical School, Boston, MA, USA. 16 Institució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Spain. 17 Present address: GENOMICS plc, Oxford, UK. 18 These authors contributed equally: Marta Guindo-Martínez, Ramon Amela. 19 These authors jointly supervised this work: Josep M. Mercader, David Torrents., Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

DRMaestro: orchestrating disaggregated resources on virtualized data-centers

  • Carneiro do Amaral, Marcelo
  • Polo Bardés, Jorda
  • Carrera Pérez, David
  • González, Nelson
  • Yang, Chih-Chieh
  • Morari, Alessandro
  • D'Amora, Bruce
  • Youssef, Alaa
  • Steinder, Malgorzata
Modern applications demand resources at an unprecedented level. In this sense, data-centers are required to scale efficiently to cope with such demand. Resource disaggregation has the potential to improve resource-efficiency by allowing the deployment of workloads in more flexible ways. Therefore, the industry is shifting towards disaggregated architectures, which enables new ways to structure hardware resources in data centers. However, determining the best performing resource provisioning is a complicated task. The optimality of resource allocation in a disaggregated data center depends on its topology and the workload collocation. This paper presents DRMaestro, a framework to orchestrate disaggregated resources transparently from the applications. DRMaestro uses a novel flow-network model to determine the optimal placement in multiple phases while employing best-efforts on preventing workload performance interference. We first evaluate the impact of disaggregation regarding the additional network requirements under higher network load. The results show that for some applications the impact is minimal, but other ones can suffer up to 80% slowdown in the data transfer part. After that, we evaluate DRMaestro via a real prototype on Kubernetes and a trace-driven simulation. The results show that DRMaestro can reduce the total job makespan with a speedup of up to ≈1.20x and decrease the QoS violation up to ≈2.64x comparing with another orchestrator that does not support resource disaggregation., This project is supported by the IBM/BSC Technology Center for Supercomputing collaboration agreement. It has also received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 639595). It is also partially supported by the Ministry of Economy of Spain under contract TIN2015-65316-P and Generalitat de Catalunya under contract 2014SGR1051, by the ICREA Academia program, and by the BSC-CNS Severo Ochoa program (SEV-2015-0493)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Design trade-offs for emerging HPC processors based on mobile market technology

  • Armejach Sanosa, Adrià
  • Casas, Marc
  • Moretó Planas, Miquel
This is a post-peer-review, pre-copyedit version of an article published in The Journal of Supercomputing. The final authenticated version is available online at: http://dx.doi.org/10.1007/s11227-019-02819-4, High-performance computing (HPC) is at the crossroads of a potential transition toward mobile market processor technology. Unlike in prior transitions, numerous hardware vendors and integrators will have access to state-of-the-art processor designs due to Arm’s licensing business model. This fact gives them greater flexibility to implement custom HPC-specific designs. In this paper, we undertake a study to quantify the different energy-performance trade-offs when architecting a processor based on mobile market technology. Through detailed simulations over a representative set of benchmarks, our results show that: (i) a modest amount of last-level cache per core is sufficient, leading to significant power and area savings; (ii) in-order cores offer favorable trade-offs when compared to out-of-order cores for a wide range of benchmarks; and (iii) heterogeneous configurations help to improve processor performance and energy efficiency., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

ECHOFS: a scheduler-guided temporary filesystem to leverage node-local NVMS

  • Miranda, Alberto
  • Nou Castell, Ramon
  • Cortés, Toni
© 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., The growth in data-intensive scientific applications poses strong demands on the HPC storage subsystem, as data needs to be copied from compute nodes to I/O nodes and vice versa for jobs to run. The emerging trend of adding denser, NVM-based burst buffers to compute nodes, however, offers the possibility of using these resources to build temporary file systems with specific I/O optimizations for a batch job. In this work, we present echofs, a temporary filesystem that coordinates with the job scheduler to preload a job's input files into node-local burst buffers. We present the results measured with NVM emulation, and different FS backends with DAX/FUSE on a local node, to show the benefits of our proposal and such coordination., This work was partially supported by the Spanish Ministry of Science and Innovation under the TIN2015–65316 grant, the Generalitat de Catalunya under contract 2014– SGR–1051, as well as the European Union’s Horizon 2020 Research and Innovation Programme, under Grant Agreement no. 671951 (NEXTGenIO). Source code available at https://github.com/bsc-ssrg/echofs., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Intelligent colocation of HPC workloads

  • Vieira Zacarias, Felippe
  • Petrucci, Vinicius
  • Nishtala, Rajiv
  • Carpenter, Paul Matthew
  • Mossé, Daniel
Many server applications suffer from a bottleneck in the shared caches, instruction execution units, I/O or memory bandwidth, even though the remaining resources may be underutilized. It is hard for developers and runtime systems to ensure that all critical resources are fully exploited by a single application, so an attractive technique for increasing server system utilization is to colocate multiple applications on the same server. When applications share critical resources, however, contention on shared resources may lead to reduced application performance. In this paper, we show that server efficiency can be improved by first modeling the expected performance degradation of colocated applications based on measured hardware performance counters, and then exploiting the model to determine an optimized mix of colocated applications. This paper presents a new intelligent resource manager and makes the following contributions: (1) a new machine learning model to predict the performance degradation of colocated applications based on hardware counters and (2) an intelligent scheduling scheme deployed on an existing resource manager to enable application co-scheduling with minimum performance degradation. Our results show that our approach achieves performance improvements of 7 % (avg) and 12 % (max) compared to the standard policy commonly used by existing job managers., This work is part of a project that has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 754337 (EuroEXA); it has been supported by the Spanish Ministry of Science and Technology (project TIN2015-65316-P), Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272), and the Severo Ochoa Programme (SEV-2015-0493). The work is also supported by FAPESB (Edital JCB 008/2015) and the Brazilian federal government under CNPq grant (Process no 430188/2018-8)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Simulating the behavior of the human brain on GPUS

  • Valero-Lara, Pedro
  • Martinez-Perez, Ivan
  • Sirvent, Raul
  • Peña, Antonio J.
  • Martorell Bofill, Xavier
  • Labarta Mancho, Jesús José
The simulation of the behavior of the Human Brain is one of the most important challenges in computing today. The main problem consists of finding efficient ways to manipulate and compute the huge volume of data that this kind of simulations need, using the current technology. In this sense, this work is focused on one of the main steps of such simulation, which consists of computing the Voltage on neurons’ morphology. This is carried out using the Hines Algorithm and, although this algorithm is the optimum method in terms of number of operations, it is in need of non-trivial modifications to be efficiently parallelized on GPUs. We proposed several optimizations to accelerate this algorithm on GPU-based architectures, exploring the limitations of both, method and architecture, to be able to solve efficiently a high number of Hines systems (neurons). Each of the optimizations are deeply analyzed and described. Two different approaches are studied, one for mono-morphology simulations (batch of neurons with the same shape) and one for multi-morphology simulations (batch of neurons where every neuron has a different shape). In mono-morphology simulations we obtain a good performance using just a single kernel to compute all the neurons. However this turns out to be inefficient on multi-morphology simulations. Unlike the previous scenario, in multi-morphology simulations a much more complex implementation is necessary to obtain a good performance. In this case, we must execute more than one single GPU kernel. In every execution (kernel call) one specific part of the batch of the neurons is solved. These parts can be seen as multiple and independent tridiagonal systems. Although the present paper is focused on the simulation of the behavior of the Human Brain, some of these techniques, in particular those related to the solving of tridiagonal systems, can be also used for multiple oil and gas simulations. Our studies have proven that the optimizations proposed in the present work can achieve high performance on those computations with a high number of neurons, being our GPU implementations about 4× and 8× faster than the OpenMP multicore implementation (16 cores), using one and two NVIDIA K80 GPUs respectively. Also, it is important to highlight that these optimizations can continue scaling, even when dealing with a very high number of neurons., This project has received funding from the European Union’s Horizon 2020 Research and Innovation Programme under Grant Agreement No. 720270 (HBP SGA1), from the Spanish Ministry of Economy and Competitiveness under the project Computación de Altas Prestaciones VII (TIN2015-65316-P), the Departament d’Innovació, Universitats i Empresa de la Generalitat de Catalunya, under project MPEXPAR: Models de Programació i Entorns d’Execució Parallels (2014-SGR-1051). We thank the support of NVIDIA through the BSC/UPC NVIDIA GPU Center of Excellence, and the European Union’s Horizon 2020 Research and Innovation Program under the Marie Sklodowska-Curie Grant Agreement No. 749516., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

sLASs: a fully automatic auto-tuned linear algebra library based on OpenMP extensions implemented in OmpSs (LASs Library)

  • Valero Lara, Pedro
  • Catalán Pallarés, Sandra
  • Martorell Bofill, Xavier
  • Usui, Tetsuzo
  • Labarta Mancho, Jesús José
© 2019 Elsevier. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/, In this work we have implemented a novel Linear Algebra Library on top of the task-based runtime OmpSs-2. We have used some of the most advanced OmpSs-2 features; weak dependencies and regions, together with the final clause for the implementation of auto-tunable code for the BLAS-3 trsm routine and the LAPACK routines npgetrf and npgesv. All these implementations are part of the first prototype of sLASs library, a novel library for auto-tunable codes for linear algebra operations based on LASs library. In all these cases, the use of the OmpSs-2 features presents an improvement in terms of execution time against other reference libraries such as, the original LASs library, PLASMA, ATLAS and Intel MKL. These codes are able to reduce the execution time in about 18% on big matrices, by increasing the IPC on gemm and reducing the time of task instantiation. For a few medium matrices, benefits are also seen. For small matrices and a subset of medium matrices, specific optimizations that allow to increase the degree of parallelism in both, gemm and trsm tasks, are applied. This strategy achieves an increment in performance of up to 40%., This project has received funding from the Spanish Ministry of Economy and Competitiveness, Spain under the project Computación de Altas Prestaciones VII (TIN2015- 65316-P), the Departament d’Innovació, Universitats i Empresa de la Generalitat de Catalunya, Spain, under project MPEXPAR: Models de Programació i Entorns d’Execució Parallels (2014-SGR-1051), and the Juan de la Cierva Grant Agreement No IJCI-2017- 33511. We also acknowledge the funding provided by Fujitsu, Japan under the BSC-Fujitsu joint project: Math Libraries Migration and Optimization., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Freezing time emulating new and faster devices with virtual machines

  • Bona, Luis C.E.
  • Elias, Alessandro
  • Ziviani, Andre P.
  • Nou, Ramon
  • Cortés, Toni
  • Alves, Marco A.Z.
Recent proposals of emerging data storage devices make it necessary to reevaluate all levels of the storage hierarchy to optimize the software stack performance. However, these new devices are not always widely available and therefore early experiments may be impossible. Emulators aim at mimicking as close as possible the behavior of a component, nonetheless, emulating new and fast storage devices is a challenging task due to time perception. In this work, we propose an approach to emulate storage devices using virtual machines (VMs) allowing the evaluation of a new device within a real system. We use a technique called freezing time, which pauses a VM to manipulate its clock and hide the real I/O completion time. Our approach is implemented at the hypervisor level and it is transparent to the guest operating system or application. We evaluate the technique under a real system using regular magnetic disks to emulate faster storage devices. Our method presented a latency error of 6.5% compared to a real device. Moreover, decoupled experiment between two laboratories, at the Barcelona Super Computing Center (BSC) in Spain, and the Center of Computer Science and Free Software (C3SL) in Brazil, demonstrated that our approach is reproducible and promising to allow the virtual evaluation of next-gen storage devices., This work was partially supported by the Spanish Ministry of Science and Innovation under the TIN2015-65316 Grant, the Generalitat de Catalunya under contract 2014-SGR-1051, the Serrapilheira Institute (Grant number Serra-1709-16621), as well as the European Union’s Horizon 2020 Research and Innovation Programme, under Grant Agreement no. 671951 (NEXTGenIO) for the extensions added after the MASCOTS paper., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

An architecture model for a distributed virtualization system

  • Pessolani, Pablo
  • Tinetti, Fernando
  • Cortés, Toni
  • Gonnet, Silvio
This article presents an architecture model for a Distributed Virtualization System, which could expand a virtual execution environment from a single physical machine to several nodes of a cluster. With current virtualization technologies, computing power and resource usage of Virtual Machines (or Containers) are limited to the physical machine where they run. To deliver high levels of performance and scalability, cloud applications are usually partitioned in several Virtual Machines (or Containers) located on different nodes of a virtualization cluster. Developers often use that processing model because the same instance of the operating system is not available on each node where their components run. The proposed architecture model is suitable for new trends in software development because it is inherently distributed. It combines and integrates Virtualization and Distributed Operating Systems technologies with the benefits of both worlds, providing the same isolated instance of a Virtual Operating System on each cluster node. Although it requires the introduction of changes in existing operating systems, thousands of legacy applications would not require modifications to obtain their benefits. A Distributed Virtualization System is suitable to deliver high-performance cloud services with provider-class features, such as high-availability, replication, migration, and load balancing. Furthermore, it is able to concurrently run several isolated instances of different guest Virtual Operating Systems, allocating a subset of nodes for each instance and sharing nodes between them. Currently, a prototype is running on a cluster of commodity hardware provided with two kinds of Virtual Operating Systems tailored for internet services (web server) as a proof of concept., Toni Cortes’ involvement in this work was financially supported by the Spanish National Government (financial grant SEV2015-0493 of Severo Ochoa program), the Spanish Ministry of Science and Innovation (contract TIN2015-65316), and the Government of Catalonia (contract 2014-SGR-1051).Fernando G. Tinetti’s involvement in this work was financially supported by Universidad de La Plata (Faculty of Informatics) and the Scientific Research Board (CIC, for its Spanish initials) of Buenos Aires, Argentina., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Main memory latency simulation: the missing link

  • Sánchez Verdejo, Rommel
  • Asifuzzaman, Kazi
  • Radulović, Milan
  • Radojković, Petar
  • Ayguadé Parra, Eduard
  • Jacob, Bruce
The community accepted the need for a detailed simulation of main memory. Currently, the CPU simulators are usually coupled with the cycle-accurate main memory simulators. However, coupling CPU and memory simulators is not a straight-forward task because some pieces of the circuitry between the last level cache and the memory DIMMs could be easily overlooked and therefore not accounted for. In this paper, we take an approach to quantify the missing cycles in the main memory simulation. To that end, we execute a memory intensive microbenchmark to validate a simulation infrastructure based on ZSim and DRAMsim2 modeling an Intel Sandy Bridge E5-2670 system. We execute the same microbenchmark on a real Sandy Bridge E5-2670 machine identifying a missing 20 ns in the simulator measurements. This is a huge difference that, in the system under study, corresponds to one-third of the overall main memory latency. We propose multiple schemes to add an extra delay in the simulation model to account for the missing cycles. Furthermore, we validate the proposals using the SPEC CPU2006 benchmarks. Finally, we repeat the main memory latency measurements on seven mainstream and emerging computing platforms. Our results show that latency between the Last Level Cache (LLC) and the main memory ranges between tens and hundreds of nanoseconds, so we emphasize on properly adjust and validate these parameters in system simulators before any measurements are performed. Overall, we believe this study would improve main memory simulation leading to the better overall system analysis and explorations performed in the computer architecture community., This work was supported by the Collaboration Agreement between Samsung Electronics Co. Ltd. and BSC, Spanish Ministry of Science and Technology (project TIN2015-65316-P), Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272) and the Severo Ochoa Programme (SEV-2015-0493) of the Spanish Government., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Ad hoc file systems for high-performance computing

  • Brinkmann, Andre
  • Mohror, Kathryn
  • Yu, Weikuan
  • Carns, Philip
  • Cortés, Toni
  • Klasky, Scott A.
  • Miranda Bueno, Alberto
  • Pfreundt, Franz-Josef
  • Ross, Robert B.
  • Vef, Marc-André
Storage backends of parallel compute clusters are still based mostly on magnetic disks, while newer and faster storage technologies such as flash-based SSDs or non-volatile random access memory (NVRAM) are deployed within compute nodes. Including these new storage technologies into scientific workflows is unfortunately today a mostly manual task, and most scientists therefore do not take advantage of the faster storage media. One approach to systematically include nodelocal SSDs or NVRAMs into scientific workflows is to deploy ad hoc file systems over a set of compute nodes, which serve as temporary storage systems for single applications or longer-running campaigns. This paper presents results from the Dagstuhl Seminar 17202 “Challenges and Opportunities of User-Level File Systems for HPC” and discusses application scenarios as well as design strategies for ad hoc file systems using node-local storage media. The discussion includes open research questions, such as how to couple ad hoc file systems with the batch scheduling environment and how to schedule stage-in and stage-out processes of data between the storage backend and the ad hoc file systems. Also presented are strategies to build ad hoc file systems by using reusable components for networking and how to improve storage device compatibility. Various interfaces and semantics are presented, for example those used by the three ad hoc file systems BeeOND, GekkoFS, and BurstFS. Their presentation covers a range from file systems running in production to cutting-edge research focusing on reaching the performance limits of the underlying devices., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

A visual embedding for the unsupervised extraction of abstract semantics

  • Garcia Gasulla, Dario
  • Ayguadé Parra, Eduard
  • Labarta Mancho, Jesús José
  • Béjar Alonso, Javier
  • Cortés García, Claudio Ulises
  • Suzumura, Toyotaro
  • Chen, R
Vector-space word representations obtained from neural network models have been shown to enable semantic operations based on vector arithmetic. In this paper, we explore the existence of similar information on vector representations of images. For that purpose we define a methodology to obtain large, sparse vector representations of image classes, and generate vectors through the state-of-the-art deep learning architecture GoogLeNet for 20 K images obtained from ImageNet. We first evaluate the resultant vector-space semantics through its correlation with WordNet distances, and find vector distances to be strongly correlated with linguistic semantics. We then explore the location of images within the vector space, finding elements close in WordNet to be clustered together, regardless of significant visual variances (e.g., 118 dog types). More surprisingly, we find that the space unsupervisedly separates complex classes without prior knowledge (e.g., living things). Afterwards, we consider vector arithmetics. Although we are unable to obtain meaningful results on this regard, we discuss the various problem we encountered, and how we consider to solve them. Finally, we discuss the impact of our research for cognitive systems, focusing on the role of the architecture being used., This work is partially supported by the Joint Study Agreement no. W156463 under the IBM/BSC Deep Learning Center agreement, by the Spanish Government through Programa Severo Ochoa (SEV-2015-0493), by the Spanish Ministry of Science and Technology through TIN2015-65316-P project and by the Generalitat de Catalunya (contracts 2014-SGR-1051), and by the Core Research for Evolutional Science and Technology (CREST) program of Japan Science and Technology Agency (JST)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Constant-time approximate sliding window framework with error control

  • Villalba Navarro, Álvaro
  • Carrera Pérez, David
Stream Processing is a crucial element for the Edge Computing paradigm, in which large amount of devices generate data at the edge of the network. This data needs to be aggregated and processed on-the-move across different layers before reaching the Cloud. Therefore, defining Stream Processing services that adapt to different levels of resource availability is of paramount importance. In this context, Stream Processing frameworks need to combine efficient algorithms with low computational complexity to manage sliding windows, with the ability to adjust resource demands for different deployment scenarios, from very low capacity edge devices to virtually unlimited Cloud platforms. The Approximate Computing paradigm provides improved performance and adaptive resource demands in data analytics, at the price of introducing some level of inaccuracy that can be calculated. In this paper we present the Approximate and Amortized Monoid Tree Aggregator (A 2 MTA). It is, to our knowledge, the first general purpose sliding window programable framework that combines constant-time aggregations with error bounded approximate computing techniques. It is very suitable for adverse stream processing environments, such as resource scarce multi-tenant edge computing. The framework can compute aggregations over multiple data dimensions, setting error bounds on any of them, and has been designed to support decoupling computation and data storage through the use of distributed Key-Value Stores to keep window elements and partial aggregations., This project is partially supported by the European Research Council (ERC), Spain under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 639595). It is also partially supported by the Ministry of Economy of Spain under contract TIN2015-65316-P and Generalitat de Catalunya, Spain under contract 2014SGR1051, by the ICREA Academia program, and by the BSC-CNS Severo Ochoa program (SEV-2015-0493)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Enabling genomics pipelines in commodity personal computers with flash storage

  • Cadenelli, Nicola
  • Jun, Sang-Woo
  • Polo Bardés, Jorda
  • Wright, Andrew J.
  • Carrera Pérez, David
  • Mithal, Arvind
Analysis of a patient’s genomics data is the first step toward precision medicine. Such analyses are performed on expensive enterprise-class server machines because input data sets are large, and the intermediate data structures are even larger (TB-size) and require random accesses. We present a general method to perform a specific genomics problem, mutation detection, on a cheap commodity personal computer (PC) with a small amount of DRAM. We construct and access large histograms of k-mers efficiently on external storage (SSDs) and apply our technique to a state-of-the-art referencefree genomics algorithm, SMUFIN, to create SMUFIN-F. We show that on two PCs, SMUFIN-F can achieve the same throughput at only one third (36%) the hardware cost and half (45%) the energy compared to SMUFIN on an enterprise-class server. To the best of our knowledge, SMUFIN-F is the first reference-free system that can detect somatic mutations on commodity PCs for whole human genomes. We believe our technique should apply to other k-mer or n-gram-based algorithms., This work was supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreements No. 639595); the Ministry of Economy of Spain under contract TIN2015-65316-P and Generalitat de Catalunya under contract 2014SGR1051; the ICREA Academia program; the BSC-CNS Severo Ochoa program (SEV-2015-0493); MIT via the NSF grant (CCF1725303); and travel support was provided by the MIT-Spain la Caixa Foundation Seed Fund., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Task-based programming in COMPSs to converge from HPC to big data

  • Conejero, Javier
  • Corella, Sandra
  • Badia Sala, Rosa Maria
  • Labarta Mancho, Jesús José
Task-based programming has proven to be a suitable model for high-performance computing (HPC) applications. Different implementations have been good demonstrators of this fact and have promoted the acceptance of task-based programming in the OpenMP standard. Furthermore, in recent years, Apache Spark has gained wide popularity in business and research environments as a programming model for addressing emerging big data problems. COMP Superscalar (COMPSs) is a task-based environment that tackles distributed computing (including Clouds) and is a good alternative for a task-based programming model for big data applications. This article describes why we consider that task-based programming models are a good approach for big data applications. The article includes a comparison of Spark and COMPSs in terms of architecture, programming model, and performance. It focuses on the differences that both frameworks have in structural terms, on their programmability interface, and in terms of their efficiency by means of three widely known benchmarking kernels: Wordcount, Kmeans, and Terasort. These kernels enable the evaluation of the more important functionalities of both programming models and analyze different work flows and conditions. The main results achieved from this comparison are (1) COMPSs is able to extract the inherent parallelism from the user code with minimal coding effort as opposed to Spark, which requires the existing algorithms to be adapted and rewritten by explicitly using their predefined functions, (2) it is an improvement in terms of performance when compared with Spark, and (3) COMPSs has shown to scale better than Spark in most cases. Finally, we discuss the advantages and disadvantages of both frameworks, highlighting the differences that make them unique, thereby helping to choose the right framework for each particular objective., This work is supported by the Spanish Government (SEV2015-0493), by the Spanish Ministry of Science and Innovation (contract TIN2015-65316-P), by Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272). Javier Conejero’s postdoctoral contract is cofinanced by the Ministry of Economy and Competitiveness under the Juan de la Cierva Formación postdoctoral fellowship number FJCI-2015-24651. This work is also supported by the Intel-BSC Exascale Lab. The Human Brain Project receives funding from the EU’s Seventh Framework Programme (FP7/2007-2013) under grant agreement no 604102., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Semi-automatic validation of cycle-accurate simulation infrastructures: The case for gem5-x86

  • Cebrián González, Juan Manuel
  • Barredo Ferreira, Adrián
  • Caminal Pallarés, Helena
  • Moretó Planas, Miquel
  • Casas, Marc
  • Valero Cortés, Mateo
Since the early 70s, simulation infrastructures have been a keystone in computer architecture research, providing a fast and reliable way to prototype and evaluate ideas for future computing systems. There are different types of simulators, from most detailed (cycle-accurate) to time-based/functional and analytical modeling. Increasing accuracy translates into several orders of magnitude in terms of simulation speed. Yet, a question remains open: are the results derived from the simulation infrastructure representative of a real machine? Validation of these infrastructures is complex and costly, usually performed upon release. However, most simulators do not provide the appropriate means to verify or validate new architectural models. In this paper, we introduce a semi-automatic validation framework based on real-hardware performance counter information. The framework provides two levels of abstraction: (a) a high level definition of the processor behavior (Top-Down model) and (b) detailed per-structure and per-pipeline-stage usage breakdown to pinpoint simulator issues. We used this framework to validate the latest available gem5-x86 simulation environment, and found several sources of error that alter the expected behavior of the simulated processor, which we were later to document and correct., This work has been partially supported by the Spanish Government (Severo Ochoa grants SEV2015-0493, SEV-2011-00067), the Spanish Ministry of Science and Innovation (contract TIN2015- 65316-P), Generalitat de Catalunya, Spain (contracts 2014-SGR1051 and 2014-SGR-1272), the RoMoL ERC Advanced, Spain Grant (GA 321253), the European HiPEAC Network of Excellence and the Mont-Blanc project (EU-FP7-610402 and EU-H2020-779877). A. Barredo has been supported by the Spanish Government under Formación del Personal Investigador fellowship number BES2017-080635. M. Moreto and M. Casas have been partially supported by the Spanish Ministry of Economy, Industry and Competitiveness, Spain under Ramon y Cajal fellowship numbers RYC2016-21104 and RYC-2017-23269., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Tailwind: Fast and atomic RDMA-based replication

  • Taleb, Yacine
  • Stutsman, Ryan
  • Antoniu, Gabriel
  • Cortés, Toni
Replication is essential for fault-tolerance. However, in in-memory systems, it is a source of high overhead. Remote direct memory access (RDMA) is attractive to create redundant copies of data, since it is low-latency and has no CPU overhead at the target. However, existing approaches still result in redundant data copying and active receivers. To ensure atomic data transfers, receivers check and apply only fully received messages. Tailwind is a zero-copy recovery-log replication protocol for scale-out in-memory databases. Tailwind is the first replication protocol that eliminates all CPU-driven data copying and fully bypasses target server CPUs, thus leaving backups idle. Tailwind ensures all writes are atomic by leveraging a protocol that detects incomplete RDMA transfers. Tailwind substantially improves replication throughput and response latency compared with conventional RPC-based replication. In symmetric systems where servers both serve requests and act as replicas, Tailwind also improves normal-case throughput by freeing server CPU resources for request processing. We implemented and evaluated Tailwind on RAMCloud, a low-latency in-memory storage system. Experiments show Tailwind improves RAMCloud’s normal-case request processing throughput by 1.7×. It also cuts down writes median and 99th percentile latencies by 2x and 3x respectively., This work has been supported by the BigStorage project, funded by the European Union under the Marie SklodowskaCurie Actions (H2020-MSCA-ITN-2014-642963), by the Spanish Ministry of Science and Innovation (contract TIN2015- 65316), by Generalitat de Catalunya (contract 2014- SGR-1051). This material is based upon work supported by the National Science Foundation under Grant Nos. CNS-1566175 and CNS-1750558. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. This work was supported in part by Facebook and VMware., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Building graph representations of deep vector embeddings

  • Garcia Gasulla, Dario
  • Vilalta Arias, Armand
  • Parés Pont, Ferran
  • Moreno Vázquez, Jonatan
  • Ayguadé Parra, Eduard
  • Labarta Mancho, Jesús José
  • Cortés García, Claudio Ulises
  • Suzumura, Toyotaro
Patterns stored within pre-trained deep neural networks compose large and powerful descriptive languages that can be used for many different purposes. Typically, deep network representations are implemented within vector embedding spaces, which enables the use of traditional machine learning algorithms on top of them. In this short paper we propose the construction of a graph embedding space instead, introducing a methodology to transform the knowledge coded within a deep convolutional network into a topological space (i.e. a network). We outline how such graph can hold data instances, data features, relations between instances and features, and relations among features. Finally, we introduce some preliminary experiments to illustrate how the resultant graph embedding space can be exploited through graph analytics algorithms, This work is partially supported by the Joint Study Agreement no. W156463 under the IBM/BSC Deep Learning Center agreement, by the Spanish Government through Programa Severo Ochoa (SEV-2015- 0493), by the Spanish Ministry of Science and Technology through TIN2015-65316-P project, by the Generalitat de Catalunya (contracts 2014-SGR-1051), and by the Core Research for Evolutional Science and Technology (CREST) program of Japan Science and Technology Agency (JST)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Artificial neural networks as emerging tools for earthquake detection

  • Rojas, Otilio
  • Otero Calviño, Beatriz
  • Alvarado, Leonardo
  • Mus, Sergi
  • Tous Liesa, Rubén
As seismic networks continue to spread and monitoring sensors become more ef¿cient, the abundance of data highly surpasses the processing capabilities of earthquake interpretation analysts. Earthquake catalogs are fundamental for fault system studies, event modellings, seismic hazard assessment, forecasting, and ultimately, for mitigating the seismic risk. These have fueled the research for the automation of interpretation tasks such as event detection, event identi¿cation, hypocenter location, and source mechanism analysis. Over the last forty years, traditional algorithms based on quantitative analyses of seismic traces in the time or frequency domain, have been developed to assist interpretation. Alternatively, recentadvancesarerelatedtotheapplicationofArti¿cial Neural Networks (ANNs), a subset of machine learning techniques that is pushing the state-of-the-art forward in many areas. Appropriated trained ANN can mimic the interpretation abilities of best human analysts, avoiding the individual weaknesses of most traditional algorithms, and spending modest computational resources at the operational stage. In this paper, we will survey the latest ANN applications to the automatic interpretation of seismic data, with a special focus on earthquake detection, and the estimation of onset times. For a comparative framework, we give an insight into the labor of human interpreters, who may face uncertainties in the case of small magnitude earthquakes., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

MPI+X: task-based parallelisation and dynamic load balance of finite element assembly

  • Garcia, Marta
  • Houzeaux, Guillaume
  • Ferrer, Roger
  • Artigues, Antoni
  • López, Victor
  • Labarta Mancho, Jesús José
  • Vázquez, Mariano
The main computing phases of numerical methods for solving partial differential equations are the algebraic system assembly and the iterative solver. This work focuses on the first task, in the context of a hybrid MPI+X paradigm. The matrix assembly consists of a loop over the elements, faces, edges or nodes of the MPI partitions to compute element matrices and vectors and then of their assemblies. In a MPI+X hybrid parallelism context, X has consisted traditionally of loop parallelism using OpenMP, with different techniques to avoid the race condition, but presenting efficiency or implementation drawbacks. We propose an alternative, based on task parallelism using some extensions to the OpenMP programming model. In addition, dynamic load balance will be applied, especially efficient in the presence of hybrid meshes. This paper presents the proposed methodology, its implementation and its validation through the solution of large computational mechanics problems up to 16k cores., The research leading to these results has received funding from: the European Union's Horizon 2020 Programme (2014–2020) and from Brazilian Ministry of Science, Technology and Innovation through Rede Nacional de Pesquisa (RNP), HPC4E Project, grant agreement 689772; the Energy oriented Centre of Excellence (EoCoE), grant agreement number 676629, funded within the Horizon2020 framework of the European Union; The Spanish Government (grant SEV2015-0493 of the Severo Ochoa Program); the Spanish Ministry of Science and Innovation (contract TIN2015-65316-P); the Generalitat de Catalunya (contract 2014-SGR-1051); the Intel-BSC Exascale Lab collaboration project. Comissió Interdepartamental de Recerca i Innovació Tecnológica(Interdepartmental Commission for Research and Technological Innovation), Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Mainstream vs. emerging HPC: metrics, trade-offs and lessons learned

  • Radulović, Milan
  • Asifuzzaman, Kazi
  • Živanovič, Darko
  • Rajovic, Nikola
  • Colin de Verdiére, Guillaume
  • Pleiter, Dirk
  • Marazakis, Manolis
  • Kallimanis, Nikolaos
  • Carpenter, Paul Matthew
  • Radojković, Petar
  • Ayguadé Parra, Eduard
©2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Various servers with different characteristics and architectures are hitting the market, and their evaluation and comparison in terms of HPC features is complex and multidimensional. In this paper, we share our experience of evaluating a diverse set of HPC systems, consisting of three mainstream and five emerging architectures. We evaluate the performance and power efficiency using prominent HPC benchmarks, High-Performance Linpack (HPL) and High Performance Conjugate Gradients (HPCG), and expand our analysis using publicly available specialized kernel benchmarks, targeting specific system components. In addition to a large body of quantitative results, we emphasize six usually overlooked aspects of the HPC platforms evaluation, and share our conclusions and lessons learned. Overall, we believe that this paper will improve the evaluation and comparison of HPC platforms, making a first step towards a more reliable and uniform methodology., This work was supported by the Spanish Ministry of Science and Technology (project TIN2015-65316-P), Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272), Severo Ochoa Programme (SEV-2015-0493) of the Spanish Government; and the European Union’s Horizon 2020 research and innovation programme under ExaNoDe project (grant agreement No 671578)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Multi-GPU parallelization of the NAS multi-zone parallel benchmarks

  • González Tallada, Marc
  • Morancho Llena, Enrique
GPU-based computing systems have become a widely accepted solution for the high-performance-computing (HPC) domain. GPUs have shown highly competitive performance-per-watt ratios and can exploit an astonishing level of parallelism. However, exploiting the peak performance of such devices is a challenge, mainly due to the combination of two essential aspects of multi-GPU execution. On one hand, the workload should be distributed evenly among the GPUs. On the other hand, communications between GPU devices are costly and should be minimized. Therefore, a trade-of between work-distribution schemes and communication overheads will condition the overall performance of parallel applications run on multi-GPU systems. In this article we present a multi-GPU implementation of NAS Multi-Zone Parallel Benchmarks (which execution alternate communication and computational phases). We propose several work-distribution strategies that try to evenly distribute the workload among the GPUs. Our evaluations show that performance is highly sensitive to this distribution strategy, as the the communication phases of the applications are heavily affected by the work-distribution schemes applied in computational phases. In particular, we consider Static, Dynamic, and Guided schedulers to find a trade-off between both phases to maximize the overall performance. In addition, we compare those schedulers with an optimal scheduler computed offline using IBM CPLEX. On an evaluation environment composed of 2 x IBM Power9 8335-GTH and 4 x GPU NVIDIA V100 (Volta), our multi-GPU parallelization outperforms single-GPU execution from 1.48x to 1.86x (2 GPUs) and from 1.75x to 3.54x (4 GPUs). This article analyses these improvements in terms of the relationship between the computational and communication phases of the applications as the number of GPUs is increased. We prove that Guided schedulers perform at similar level as optimal schedulers., This work was supported by the Spanish Ministry of Science and Technology (TIN2015-65316-P) and by the Generalitat de Catalunya (2014-SGR-1051)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Workload-aware placement strategies to leverage disaggregated resources in the datacenter

  • Call Barreiro, Erin
  • Polo Bardés, Jorda
  • Carrera Pérez, David
Disaggregation of resources is a datacenter strategy that aims to decouple the physical location of resources from the place where they are accessed, as opposed to physically attached devices connected to the Peripheral Component Interconnect Express (PCIe) bus. By attaching and detaching resources through a fast interconnection network, it is possible to increase the flexibility to manage datacenter infrastructures while keeping the performance of the pooled and disaggregated devices. This article introduces workload scheduling and placement policies for environments with disaggregated memories. These policies are driven by accurate prebuilt performance degradation models. We focus on the use of nonvolatile memory to store data and/or to provide memory extensions. Following a software-defined approach, persistent memories are combined to provide higher capacity and/or bandwidth devices, or used by multiple workloads to increase the number of running workloads. Different combinations of workloads and associated soft deadlines are used to evaluate the placement policies using a system simulator. When using the first-fit policy, results show that a disaggregated system can reduce missed deadlines up to 49% when compared to a physically attached system. When our proposed policy with workload awareness is enabled in a disaggregated system, missed deadlines can be reduced up to 100% (no deadlines missed)., This work was supported in part by the Ministry of Economy of Spain under Contract TIN2015-65316-P, in part by the Ministry of Science under Contract PID2019-107255GBC21/AEI/10.13039/501100011033, and in part by the Generalitat de Catalunya under Contract 2014SGR1051., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Application acceleration on FPGAs with OmpSs@FPGA

  • Bosch, Jaume
  • Tan, Xubin
  • Filgueras Izquierdo, Antonio
  • Vidal, Miquel
  • Mateu, Marc
  • Jiménez-González, Daniel
  • Álvarez, Carlos
  • Martorell Bofill, Xavier
  • Ayguadé Parra, Eduard
  • Labarta Mancho, Jesús José
© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., OmpSs@FPGA is the flavor of OmpSs that allows offloading application functionality to FPGAs. Similarly to OpenMP, it is based on compiler directives. While the OpenMP specification also includes support for heterogeneous execution, we use OmpSs and OmpSs@FPGA as prototype implementation to develop new ideas for OpenMP. OmpSs@FPGA implements the tasking model with runtime support to automatically exploit all SMP and FPGA resources available in the execution platform. In this paper, we present the OmpSs@FPGA ecosystem, based on the Mercurium compiler and the Nanos++ runtime system. We show how the applications are transformed to run on the SMP cores and the FPGA. The application kernels defined as tasks to be accelerated, using the OmpSs directives are: 1) transformed by the compiler into kernels connected with the proper synchronization and communication ports, 2) extracted to intermediate files, 3) compiled through the FPGA vendor HLS tool, and 4) used to configure the FPGA. Our Nanos++ runtime system schedules the application tasks on the platform, being able to use the SMP cores and the FPGA accelerators at the same time. We present the evaluation of the OmpSs@FPGA environment with the Matrix Multiplication, Cholesky and N-Body benchmarks, showing the internal details of the execution, and the performance obtained on a Zynq Ultrascale+ MPSoC (up to 128x). The source code uses OmpSs@FPGA annotations and different Vivado HLS optimization directives are applied for acceleration., This work is partially supported by the European Union H2020 program through the EuroEXA project (grant 754337), and HiPEAC (GA 687698), by the Spanish Government through Programa Severo Ochoa (SEV-2015- 0493), by the Spanish Ministry of Science and Technology (TIN2015-65316-P) and the Departament d’Innovació Universitats i Empresa de la Generalitat de Catalunya, under project MPEXPAR: Models de Programació i Entorns d’Execució Paral·lels (2014-SGR-1051)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Real-time logo detection in brand-related social media images

  • Orti, Oscar
  • Tous Liesa, Rubén
  • Gómez Parada, Mauro
  • Poveda, Jonatan
  • Cruz de la Cruz, Stalin Leonel
  • Wust, Otto
This paper presents a work consisting in using deep convolutional neural networks (CNNs) for real-time logo detection in brand-related social media images. The final goal is to facilitate searching and discovering user-generated content (UGC) with potential value for digital marketing tasks. The images are captured in real time and automatically annotated with two CNNs designed for object detection, SSD InceptionV2 and Faster Atrous InceptionV4 (that provides better performance on small objects). We report experiments with 2 real brands, Estrella Damm and Futbol Club Barcelona. We examine the impact of different configurations and derive conclusions aiming to pave the way towards systematic and optimized methodologies for automatic logo detection in UGC., This work is partially supported by the Spanish Ministry of Economy and Competitivity under contract TIN2015-65316-P and by the SGR programme (2014- SGR-1051 and 2017-SGR-962) of the Catalan Government., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Heterogeneous hierarchical workflow composition

  • Yildiz, Orcun
  • Ejarque Artigas, Jorge
  • Chan, Henry
  • Sankaranarayanan, Subramanian
  • Badia Sala, Rosa Maria
  • Peterka, Tom
Workflow systems promise scientists an automated end-to-end path from hypothesis to discovery. However, expecting any single workflow system to deliver such a wide range of capabilities is impractical. A more practical solution is to compose the end-to-end workflow from more than one system. With this goal in mind, the integration of task-based and in situ workflows is explored, where the result is a hierarchical heterogeneous workflow composed of subworkflows, with different levels of the hierarchy using different programming, execution, and data models. Materials science use cases demonstrate the advantages of such heterogeneous hierarchical workflow composition., This work is a collaboration between Argonne National Laboratory and the Barcelona Supercomputing Center within the Joint Laboratory for Extreme-Scale Computing. This research is supported by the U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research, under contract number DE-AC02- 06CH11357, program manager Laura Biven, and by the Spanish Government (SEV2015-0493), by the Spanish Ministry of Science and Innovation (contract TIN2015-65316-P), by Generalitat de Catalunya (contract 2014-SGR-1051)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Performance meets programmabilty: Enabling native Python MPI tasks in PyCOMPSs

  • Elshazly, Hatem Mohamed Abdelfattah Eid
  • Lordan Gomis, Francesc
  • Ejarque Artigas, Jorge
  • Badia Sala, Rosa Maria
©2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., The increasing complexity of modern and future computing systems makes it challenging to develop applications that aim for maximum performance. Hybrid parallel programming models offer new ways to exploit the capabilities of the underlying infrastructure. However, the performance gain is sometimes accompanied by increased programming complexity. We introduce an extension to PyCOMPSs, a high-level task-based parallel programming model for Python applications, to support tasks that use MPI natively as part of the task model. Without compromising application's programmability, using Native MPI tasks in PyCOMPSs offers up to 3x improvement in total performance for compute intensive applications and up to 1.9x improvement in total performance for I/O intensive applications over sequential implementation of the tasks., This work is partially supported by the European Union through the Horizon 2020 research and innovation programme under contracts 721865 (EXPERTISE Project) and 800898 (ExaQUte project), by the Spanish Government (TIN2015- 65316-P) and the Generalitat de Catalunya (contract 2014- SGR-1051)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Accelerated execution via eager-release of dependencies in task-based workflows

  • Elshazly, Hatem Mohamed Abdelfattah Eid
  • Lordan Gomis, Francesc
  • Ejarque Artigas, Jorge
  • Badia Sala, Rosa Maria
Task-based programming models offer a flexible way to express the unstructured parallelism patterns of nowadays complex applications. This expressive capability is required to achieve maximum possible performance for applications that are executed in distributed execution platforms. In current task-based workflows, tasks are launched for execution when their data dependencies are satisfied. However, even though the data dependencies of a certain task might have been already produced, the execution of this task will be delayed until its predecessor tasks completely finish their execution. As a consequence of this approach of releasing dependencies, the amount of parallelism inherent in applications is limited and performance improvement opportunities are wasted. To mitigate this limitation, we propose an eager approach for releasing data dependencies. Following this approach, the execution of tasks will not be delayed until their predecessor tasks completely finish their execution, instead, tasks will be launched for execution as soon as their data requirements are available. Hence, more parallelism is exposed and applications can achieve higher levels of performance by overlapping the execution of tasks. Towards achieving this goal, in this paper we propose applying two changes to task-based workflow systems. First, modifying the dependency relationships of tasks to be specified not only in terms of predecessor and successor tasks but also in terms of the data that caused these dependencies. Second, triggering the release of dependencies as soon as a predecessor task generates the output data instead of having to wait until the end of the predecessor execution to release all of its dependencies. We realize this proposal using PyCOMPSs: a task-based programming model for parallelizing Python applications. Our experiments show that using an eager approach for releasing dependencies achieves more than 50% performance improvement in the total execution time as compared to the default approach of releasing dependencies., This work is partially supported by the European Union through the Horizon 2020 research and innovation programme under contracts 721865 (EXPERTISE Project) by the Spanish Government (SEV2015-0493,TIN2015-65316-P) and the Generalitat de Catalunya (contract 2014-SGR1051)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

GekkoFS: A temporary distributed file system for HPC applications

  • Vef, Marc-André
  • Moti, Nafiseh
  • Süb, Tim
  • Tocci, Tommaso
  • Nou Castell, Ramon
  • Miranda, Alberto
  • Cortés, Toni
  • Brinkmann, Andre
We present GekkoFS, a temporary, highly-scalable burst buffer file system which has been specifically optimized for new access patterns of data-intensive High-Performance Computing (HPC) applications. The file system provides relaxed POSIX semantics, only offering features which are actually required by most (not all) applications. It is able to provide scalable I/O performance and reaches millions of metadata operations already for a small number of nodes, significantly outperforming the capabilities of general-purpose parallel file systems., The work has been funded by the German Research Foundation (DFG) through the ADA-FS project as part of the Priority Programme 1648. It is also supported by the Spanish Ministry of Science and Innovation (TIN2015–65316), the Generalitat de Catalunya (2014–SGR–1051), as well as the European Union’s Horizon 2020 Research and Innovation Programme (NEXTGenIO, 671951) and the European Comission’s BigStorage project (H2020-MSCA-ITN-2014-642963). This research was conducted using the supercomputer MOGON II and services offered by the Johannes Gutenberg University Mainz., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

PyCOMPSs: Parallel computational workflows in Python

  • Tejedor, Enric
  • Becerra Fontal, Yolanda
  • Alomar, Guillem
  • Queralt Calafat, Anna
  • Badia Sala, Rosa Maria
  • Torres Viñals, Jordi
  • Cortés, Toni
  • Labarta Mancho, Jesús José
The use of the Python programming language for scientific computing has been gaining momentum in the last years. The fact that it is compact and readable and its complete set of scientific libraries are two important characteristics that favour its adoption. Nevertheless, Python still lacks a solution for easily parallelizing generic scripts on distributed infrastructures, since the current alternatives mostly require the use of APIs for message passing or are restricted to embarrassingly parallel computations. In that sense, this paper presents PyCOMPSs, a framework that facilitates the development of parallel computational workflows in Python. In this approach, the user programs her script in a sequential fashion and decorates the functions to be run as asynchronous parallel tasks. A runtime system is in charge of exploiting the inherent concurrency of the script, detecting the data dependencies between tasks and spawning them to the available resources. Furthermore, we show how this programming model can be built on top of a Big Data storage architecture, where the data stored in the backend is abstracted and accessed from the application in the form of persistent objects., This work has been supported by the following institutions: the Spanish Government with grant SEV-2011-00067 of Severo Ochoa Program and contract Computaci´on de Altas Prestaciones VI (TIN2012-34557); by the SGR programme (2014-SGR-1051) of the Catalan Government; by the project The Human Brain Project, funded by the European Commission under contract 604102; and by the Intel-BSC Exascale Lab collaboration., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

BioExcel Building Blocks, a software library for interoperable biomolecular simulation workflows

  • Andrio, Pau
  • Hospital, Adam
  • Conejero Bañón, Javier
  • Jordá, Luis
  • del Pino, Marc
  • Codo, Laia
  • Soiland Reyes, Stian
  • Goble, Carole
  • Lezzi, Daniele
  • Badia Sala, Rosa Maria
  • Orozco López, Modesto
  • Gelpi Buchaca, Josep Lluís
In the recent years, the improvement of software and hardware performance has made biomolecular simulations a mature tool for the study of biological processes. Simulation length and the size and complexity of the analyzed systems make simulations both complementary and compatible with other bioinformatics disciplines. However, the characteristics of the software packages used for simulation have prevented the adoption of the technologies accepted in other bioinformatics fields like automated deployment systems, workflow orchestration, or the use of software containers. We present here a comprehensive exercise to bring biomolecular simulations to the “bioinformatics way of working”. The exercise has led to the development of the BioExcel Building Blocks (BioBB) library. BioBB’s are built as Python wrappers to provide an interoperable architecture. BioBB’s have been integrated in a chain of usual software management tools to generate data ontologies, documentation, installation packages, software containers and ways of integration with workflow managers, that make them usable in most computational environments., We thank The Barcelona Supercomputing Center for providing all the HPC resources to launch the test executions presented in this manuscript. Te project was supported by the following grants: BioExcel Center of Excellence (Horizon 2020 Framework Programme. Grants 675728, and 823830). Elixir-Excelerate (Horizon 2020 Framework Programme. Grant 676559). Spanish National Institute for Bioinformatics (Institute of Health Carlos III. Grants PT13/0001/0019, PT13/0001/0028, PT17/0009/0001). Spanish Government, Severo Ochoa Grant SEV2015-0493, TIN2015-65316-P, Generalitat de Catalunya. 2014-SGR-1051, 2017-SGR-1110., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Revisiting active object stores: Bringing data locality to the limit with NVM

  • Barceló Cuerda, Alex
  • Queralt Calafat, Anna
  • Cortés, Toni
Object stores are widely used software stacks that achieve excellent scale-out with a well-defined interface and robust performance. However, their traditional get/put interface is unable to exploit data locality at its fullest, and limits reaching its peak performance. In particular, there is one way to improve data locality that has not yet achieved mainstream adoption: the active object store. Although there are some projects that have implemented the main idea of the active object store such as Swift’s Storlets or Ceph Object Classes, the scope of these implementations is limited. We believe that there is a huge potential for active object stores in the current status quo. Hyper-converged nodes are bringing more computing capabilities to storage nodes –and viceversa. The proliferation of non-volatile memory (NVM) technology is blurring the line between system memory (fast and scarce) and block devices (slow and abundant). More and more applications need to manage a sheer amount of data (data analytics, Big Data, Machine Learning & AI, etc.), demanding bigger clusters and more complex computations. All these elements are potential game changers that need to be evaluated in the scope of active object stores. More specifically, having NVM devices presents additional opportunities, such as in-place execution. Being able to use the NVM from within the storage system while taking advantage of in-place execution (thanks to the byte-addressable nature of the NVM), in conjunction with the computing capabilities of hyper-converged nodes, can lead to active object stores that greatly outperform their non-active counterparts. In this article we propose an active object store software stack and evaluate it on an NVM-populated node. We will show how this setup is able to reduce execution times from 10% up to more than 90% in a variety of representative application scenarios. Our discussion will focus on the active aspect of the system as well as on the implications of the memory configuration., This work has been partially supported by the Spanish Government (PID2019-107255 GB), by Generalitat de Catalunya (contract 2014-SGR-1051), and by the European Commission through the Horizon 2020 Research and Innovation program and the EuroHPC JU under contract 955558 (eFlows4HPC project). Anna Queralt is a Serra Húnter Fellow. This work has been partially supported by Intel® under the BSC-Intel collaboration on 3D XPoint™ Technology., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Sampled simulation of task-based programs

  • Grass, Thomas
  • Carlson, Trevor E.
  • Rico Carro, Alejandro
  • Ceballos, Germán
  • Ayguadé Parra, Eduard
  • Casas, Marc
  • Moretó Planas, Miquel
© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works, Sampled simulation is a mature technique for reducing simulation time of single-threaded programs. Nevertheless, current sampling techniques do not take advantage of other execution models, like task-based execution, to provide both more accurate and faster simulation. Recent multi-threaded sampling techniques assume that the workload assigned to each thread does not change across multiple executions of a program. This assumption does not hold for dynamically scheduled task-based programming models. Task-based programming models allow the programmer to specify program segments as tasks which are instantiated many times and scheduled dynamically to available threads. Due to variation in scheduling decisions, two consecutive executions on the same machine typically result in different instruction streams processed by each thread. In this paper, we propose TaskPoint, a sampled simulation technique for dynamically scheduled task-based programs. We leverage task instances as sampling units and simulate only a fraction of all task instances in detail. Between detailed simulation intervals, we employ a novel fast-forwarding mechanism for dynamically scheduled programs. We evaluate different automatic techniques for clustering task instances and show that DBSCAN clustering combined with analytical performance modeling provides the best trade-off of simulation speed and accuracy. TaskPoint is the first technique combining sampled simulation and analytical modeling and provides a new way to trade off simulation speed and accuracy. Compared to detailed simulation, TaskPoint accelerates architectural simulation with 8 simulated threads by an average factor of 220x at an average error of 0.5 percent and a maximum error of 7.9 percent., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

On the behavior of convolutional nets for feature extraction

  • Garcia Gasulla, Dario
  • Parés Pont, Ferran
  • Vilalta Arias, Armand
  • Moreno, Jonatan
  • Ayguadé Parra, Eduard
  • Labarta Mancho, Jesús José
  • Cortés García, Claudio Ulises
  • Suzumura, Toyotaro
Deep neural networks are representation learning techniques. During training, a deep net is capable of generating a descriptive language of unprecedented size and detail in machine learning. Extracting the descriptive language coded within a trained CNN model (in the case of image data), and reusing it for other purposes is a field of interest, as it provides access to the visual descriptors previously learnt by the CNN after processing millions of images, without requiring an expensive training phase. Contributions to this field (commonly known as feature representation transfer or transfer learning) have been purely empirical so far, extracting all CNN features from a single layer close to the output and testing their performance by feeding them to a classifier. This approach has provided consistent results, although its relevance is limited to classification tasks. In a completely different approach, in this paper we statistically measure the discriminative power of every single feature found within a deep CNN, when used for characterizing every class of 11 datasets. We seek to provide new insights into the behavior of CNN features, particularly the ones from convolutional layers, as this can be relevant for their application to knowledge representation and reasoning. Our results confirm that low and middle level features may behave differently to high level features, but only under certain conditions. We find that all CNN features can be used for knowledge representation purposes both by their presence or by their absence, doubling the information a single CNN feature may provide. We also study how much noise these features may include, and propose a thresholding approach to discard most of it. All these insights have a direct application to the generation of CNN embedding spaces., This work is partially supported by the Joint Study Agreement no. W156463 under the IBM/BSC Deep Learning Center agreement, by the Spanish Government through Programa Severo Ochoa (SEV-2015-0493), by the Spanish Ministry of Science and Technology through TIN2015-65316-P project and by the Generalitat de Catalunya (contracts 2014-SGR-1051), and by the Core Research for Evolutional Science and Technology (CREST) program of Japan Science and Technology Agency (JST)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

cuThomasBatch and cuThomasVBatch, CUDA Routines to compute batch of tridiagonal systems on NVIDIA GPUs

  • Valero Lara, Pedro
  • Martinez Pérez, Ivan
  • Sirvent, Raül
  • Martorell Bofill, Xavier
  • Peña, Antonio J.
The solving of tridiagonal systems is one of the most computationally expensive parts in many applications, so that multiple studies have explored the use of NVIDIA GPUs to accelerate such computation. However, these studies have mainly focused on using parallel algorithms to compute such systems, which can efficiently exploit the shared memory and are able to saturate the GPUs capacity with a low number of systems, presenting a poor scalability when dealing with a relatively high number of systems. The gtsvStridedBatch routine in the cuSPARSE NVIDIA package is one of these examples, which is used as reference in this article. We propose a new implementation (cuThomasBatch) based on the Thomas algorithm. Unlike other algorithms, the Thomas algorithm is sequential, and so a coarse-grained approach is implemented where one CUDA thread solves a complete tridiagonal system instead of one CUDA block as in gtsvStridedBatch. To achieve a good scalability using this approach, it is necessary to carry out a transformation in the way that the inputs are stored in memory to exploit coalescence (contiguous threads access to contiguous memory locations). Different variants regarding the transformation of the data are explored in detail. We also explore some variants for the case of variable batch, when the size of the systems of the batch has different size (cuThomasVBatch). The results given in this study prove that the implementations carried out in this work are able to beat the reference code, being up to 5× (in double precision) and 6× (in single precision) faster using the latest NVIDIA GPU architecture, the Pascal P100., This project was funded from the European Union's Horizon 2020 research and innovation programme under grant agreement 720270 (HBPSGA1), from the Spanish Ministry of Economy and Competitiveness under the project Computación de Altas Prestaciones VII (TIN2015-65316-P)and the Departament d'Innovació, Universitats i Empresa de la Generalitat de Catalunya, under project MPEXPAR: Models de Programació iEntorns d'Execució Paral·lels (2014-SGR-1051). We thank the support of NVIDIA through the BSC/UPC NVIDIA GPU Center of Excellence andthe valuable feedback provided by Lung Sheng Chien and Alex Fit-Florea. Antonio J. Peña was cofinanced by the Spanish Ministry of Economy andCompetitiveness under Juan de la Cierva fellowship number IJCI-2015-23266., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Go with the flow: Recurrent networks for wind time series multi-step forecasting

  • Manero Font, Jaume
  • Béjar Alonso, Javier
  • Cortés García, Claudio Ulises
One of the ways of reducing the effects of Climate Change is to rely on renewable energy sources. Their intermittent nature makes necessary to obtain a mid-long term accurate forecasting. Wind Energy prediction is based on the ability to forecast wind speed. This has been a problem approached using different methods based on the statistical properties of the wind time series. Wind Time series are non-linear and non-stationary, making their forecasting very challenging. Deep neural networks have shown their success recently for problems involving sequences with non-linear behavior. In this work, we perform experiments comparing the capability of different neural network architectures for multi-step forecasting obtaining a 12 hours ahead prediction using data from the National Renewable Energy Laboratory's WIND dataset, Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

You only run once: Spark auto-tuning from a single run

  • Buchaca Prats, David
  • Albuquerque Portella, Felipe
  • Costa, Carlos H. A.
  • Berral García, Josep Lluís
Tuning configurations of Spark jobs is not a trivial task. State-of-the-art auto-tuning systems are based on iteratively running workloads with different configurations. During the optimization process, the relevant features are explored to find good solutions. Many optimizers enhance the time-to-solution using black-box optimization algorithms that do not take into account any information from the Spark workloads. In this article, we present a new method for tuning configurations that uses information from one run of a Spark workload. To achieve good performance, we mine the SparkEventLog that is generated by the Spark engine. This log file contains a large amount of information from the executed application. We use this information to enhance a performance model with low-level features from the workload to be optimized. These features include Spark Actions, Transformations, and Task metrics. This process allows us to obtain application-specific workload information. With this information our system can predict sensible Spark configurations for unseen jobs, given that it has been trained with reasonable coverage of Spark applications. Experiments show that the presented system correctly produces good configurations, while achieving up to 80% speedup with respect to the default Spark configuration, and up to 12x speedup of the time-to-solution with respect to a standard Bayesian Optimization procedure., This work is supported by the European Research Council under the European Union’s Horizon 2020 research and innovation programme (grant agreement no 639595); the Spanish Ministry of Economy, under contract TIN2015- 65316-P, and the Generalitat de Catalunya under contract 2014SGR1051; the ICREA Academia program; the BSC-CNS Severo Ochoa program (SEV-2015-0493); and by Petroleo Brasileiro S. A. (PETROBRAS)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Distributed training of deep neural networks with spark: The MareNostrum experience

  • Cruz, Leonel
  • Tous Liesa, Rubén
  • Otero Calviño, Beatriz
Deployment of a distributed deep learning technology stack on a large parallel system is a very complex process, involving the integration and configuration of several layers of both, general-purpose and custom software. The details of such kind of deployments are rarely described in the literature. This paper presents the experiences observed during the deployment of a technology stack to enable deep learning workloads on MareNostrum, a petascale supercomputer. The components of a layered architecture, based on the usage of Apache Spark, are described and the performance and scalability of the resulting system is evaluated. This is followed by a discussion about the impact of different configurations including parallelism, storage and networking alternatives, and other aspects related to the execution of deep learning workloads on a traditional HPC setup. The derived conclusions should be useful to guide similarly complex deployments in the future., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

“Dust in the wind...”, deep learning application to wind energy time series forecasting

  • Manero Font, Jaume
  • Béjar Alonso, Javier
  • Cortés García, Claudio Ulises
To balance electricity production and demand, it is required to use different prediction techniques extensively. Renewable energy, due to its intermittency, increases the complexity and uncertainty of forecasting, and the resulting accuracy impacts all the different players acting around the electricity systems around the world like generators, distributors, retailers, or consumers. Wind forecasting can be done under two major approaches, using meteorological numerical prediction models or based on pure time series input. Deep learning is appearing as a new method that can be used for wind energy prediction. This work develops several deep learning architectures and shows their performance when applied to wind time series. The models have been tested with the most extensive wind dataset available, the National Renewable Laboratory Wind Toolkit, a dataset with 126,692 wind points in North America. The architectures designed are based on different approaches, Multi-Layer Perceptron Networks (MLP), Convolutional Networks (CNN), and Recurrent Networks (RNN). These deep learning architectures have been tested to obtain predictions in a 12-h ahead horizon, and the accuracy is measured with the coefficient of determination, the R² method. The application of the models to wind sites evenly distributed in the North America geography allows us to infer several conclusions on the relationships between methods, terrain, and forecasting complexity. The results show differences between the models and confirm the superior capabilities on the use of deep learning techniques for wind speed forecasting from wind time series data., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Real-time data center's telemetry reduction and reconstruction using Markov chain models

  • Baig, Shuja-ur-Rehman
  • Iqbal, Waheed
  • Berral García, Josep Lluís
  • Erradi, Abdelkarim
  • Carrera Pérez, David
Large-scale data centers are composed of thousands of servers organized in interconnected racks to offer services to users. These data centers continuously generate large amounts of telemetry data streams (e.g., hardware utilization metrics) used for multiple purposes, including resource management, workload characterization, resource utilization prediction, capacity planning, and real-time analytics. These telemetry streams require costly bandwidth utilization and storage space, particularly at medium-long term for large data centers. This paper addresses this problem by proposing and evaluating a system to efficiently reduce bandwidth and storage for telemetry data through real-time modeling using Markov chain based methods. Our proposed solution was evaluated using real telemetry datasets and compared with polynomial regression methods for reducing and reconstructing data. Experimental results show that data can be lossy compressed up to 75% for bandwidth utilization and 95.33% for storage space, with reconstruction accuracy close to 92%., This work was supported in part by the European Research Council (ERC) under the EU Horizon 2020 programme under Grant GA 639595, in part by the Spanish Ministry of Economy, Industry and Competitiveness under Grant TIN2015-65316-P and Grant IJCI2016-27485, in part by the Generalitat de Catalunya under Grant 2014-SGR-1051, in part by the University of the Punjab, Pakistan, and in part by the Qatar National Research Fund (a member of Qatar Foundation) under NPRP Grant # NPRP9-224-1-049., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

CAPre: Code-Analysis based Prefetching for Persistent object stores

  • Touma, Rizkallah
  • Queralt Calafat, Anna
  • Cortés, Toni
Data prefetching aims to improve access times to data storage systems by predicting data records that are likely to be accessed by subsequent requests and retrieving them into a memory cache before they are needed. In the case of Persistent Object Stores, previous approaches to prefetching have been based on predictions made through analysis of the store’s schema, which generates rigid predictions, or monitoring access patterns to the store while applications are executed, which introduces memory and/or computation overhead. In this paper, we present CAPre, a novel prefetching system for Persistent Object Stores based on static code analysis of object-oriented applications. CAPre generates the predictions at compile-time and does not introduce any overhead to the application execution. Moreover, CAPre is able to predict large amounts of objects that will be accessed in the near future, thus enabling the object store to perform parallel prefetching if the objects are distributed, in a much more aggressive way than in schema-based prediction algorithms. We integrate CAPre into a distributed Persistent Object Store and run a series of experiments that show that it can reduce the execution time of applications from 9% to over 50%, depending on the nature of the application and its persistent data model., This work has been supported by the European Union’s Horizon 2020 research and innovation program under the BigStorage European Training Network (ETN) (grant H2020-MSCA-ITN-2014- 642963), the Spanish Ministry of Science and Innovation (contract TIN2015-65316) and the Generalitat de Catalunya, Spain (contract 2014-SGR-1051)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Constant-time sliding window framework with reduced memory footprint and efficient bulk evictions

  • Villalba Navarro, Álvaro
  • Berral García, Josep Lluís
  • Carrera Pérez, David
The fast evolution of data analytics platforms has resulted in an increasing demand for real-time data stream processing. From Internet of Things applications to the monitoring of telemetry generated in large data centers, a common demand for currently emerging scenarios is the need to process vast amounts of data with low latencies, generally performing the analysis process as close to the data source as possible. Stream processing platforms are required to be malleable and absorb spikes generated by fluctuations of data generation rates. Data is usually produced as time series that have to be aggregated using multiple operators, being sliding windows one of the most common abstractions used to process data in real-time. To satisfy the above-mentioned demands, efficient stream processing techniques that aggregate data with minimal computational cost need to be developed. In this paper we present the Monoid Tree Aggregator general sliding window aggregation framework, which seamlessly combines the following features: amortized O(1) time complexity and a worst-case of O(log n) between insertions; it provides both a window aggregation mechanism and a window slide policy that are user programmable; the enforcement of the window sliding policy exhibits amortized O(1) computational cost for single evictions and supports bulk evictions with cost O(log n) ; and it requires a local memory space of O(log n) . The framework can compute aggregations over multiple data dimensions, and has been designed to support decoupling computation and data storage through the use of distributed Key-Value Stores to keep window elements and partial aggregations., This project is partially supported by the European Research Council (ERC), Spain under the European Unions Horizon 2020 research and innovation programme (grant agreement No 639595). It is also partially supported by the Ministry of Economy of Spain under contract TIN2015- 65316-P and Generalitat de Catalunya, Spain under contract 2014SGR1051, by the ICREA Academia program, and by the BSC-CNS Severo Ochoa program (SEV-2015-0493)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Cost-aware prediction of uncorrected DRAM errors in the field

  • Boixaderas Coderch, Isaac
  • Živanovič, Darko
  • Moré Codina, Sergi
  • Bartolomé Rodríguez, Javier
  • Vicente Dorca, David
  • Casas, Marc
  • Carpenter, Paul Matthew
  • Radojković, Petar
  • Ayguadé Parra, Eduard
This paper presents and evaluates a method to predict DRAM uncorrected errors, a leading cause of hardware failures in large-scale HPC clusters. The method uses a random forest classifier, which was trained and evaluated using error logs from two years of production of the MareNostrum 3 supercomputer. By enabling the system to take measures to mitigate node failures, our method reduces lost compute time by up to 57%, a net saving of 21,000 node–hours per year. We release all source code as open source. We also discuss and clarify aspects of methodology that are essential for a DRAM prediction method to be useful in practice. We explain why standard evaluation metrics, such as precision and recall, are insufficient, and base the evaluation on a cost–benefit analysis. This methodology can help ensure that any DRAM error predictor is clear from training bias and has a clear cost–benefit calculation., This work was supported by the Spanish Ministry of Science and Technology (project PID2019-107255GB), Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272) and the European Union’s Horizon 2020 research and innovation programme and EuroEXA project (grant agreement No 754337). Paul Carpenter and Marc Casas hold the Ramon y Cajal fellowship under contracts RYC2018-025628-I and RYC2017-23269, respectively, of the Ministry of Economy and Competitiveness of Spain., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

STT-MRAM for real-time embedded systems: performance and WCET implications

  • Asifuzzaman, Kazi
  • Fernández, Mikel
  • Radojković, Petar
  • Abella Ferrer, Jaume
  • Cazorla Almeida, Francisco Javier
STT-MRAM is an emerging non-volatile memory quickly approaching DRAM in terms of capacity, frequency and device size. Intensified efforts in STT-MRAM research by the memory manufacturers may indicate a revolution with STT-MRAM memory technology is imminent, and therefore it is essential to perform system level research to explore use-cases and identify computing domains that could benefit from this technology. Special STT-MRAM features such as intrinsic radiation hardness, non-volatility, zero stand-by power and capability to function in extreme temperatures makes it particularly suitable for aerospace, avionics and automotive applications. Such applications often have real-time requirements --- that is, certain tasks must complete within a strict deadline. Analyzing whether this deadline is met requires Worst Case Execution Time (WCET) Analysis, which is a fundamental part of evaluating any real-time system. In this study, we investigate the feasibility of using STT-MRAM in real-time embedded systems by analyzing average system performance impact and WCET implications., This work was supported by BSC, Spanish Government through Programa Severo Ochoa (SEV-2015-0493), by the Spanish Ministry of Science and Technology through TIN2015-65316-P project and by the Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272). This work has also received funding from the European Union’s Horizon 2020 research and innovation programme under ExaNoDe project (grant agreement No 671578). Jaume Abella was partially supported by the Ministry of Economy and Competitive-ness under Ramon y Cajal postdoctoral fellowship RYC-2013-14717., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Towards mobile cloud computing with single sign-on access

  • Lordan Gomis, Francesc-Josep
  • Jensen, J. K.
  • Badia Sala, Rosa Maria
This is a post-peer-review, pre-copyedit version of an article published in Journal of Grid Computing. The final authenticated version is available online at: http://dx.doi.org/10.1007/s10723-017-9413-3, The low computing power of mobile devices impedes the development of mobile applications with a heavy computing load. Mobile Cloud Computing (MCC) has emerged as the solution to this by connecting mobile devices with the “infinite” computing power of the Cloud. As mobile devices typically communicate over untrusted networks, it becomes necessary to secure the communications to avoid privacy-sensitive data breaches. This paper presents work on implementing MCC applications with secure communications. For that purpose, we built on COMPSs-Mobile, a redesigned implementation of the COMP Superscalar (COMPSs) framework aiming to MCC platorms. COMPSs-Mobile automatically exploits the parallelism inherent in an application and orchestrates its execution on loosely-coupled distributed environment. To avoid a vendor lock-in, this extension leverages on the Generic Security Services Application Program Interface (GSSAPI) (RFC2743) as a generic way to access security services to provide communications with authentication, secrecy and integrity. Besides, GSSAPI allows applications to take profit of more advanced features, such as Federated Identity or Single Sign-On, which the underlying security framework could provide. To validate the practicality of the proposal, we use Kerberos as the security services provider to implement SSO; however, applications do not authenticate themselves and require users to obtain and place the credentials beforehand. To evaluate the performance, we conducted some tests running an application on a smartphone offloading tasks to a private cloud. Our results show that the overhead of securing the communications is acceptable., This work has been supported by the Spanish Government (contracts TIN2012-34557, TIN2015-65316-P and grants BES-2013-067167, EEBB-I-15-09808 of the Research Training Program and SEV-2011-00067 of Severo Ochoa Program), by Generalitat de Catalunya (contract 2014-SGR-1051) and by the European Commission (ASCETiC project, FP7-ICT-2013.1.2 contract 610874). The second author was partially supported by the European Commission's Horizon2020 programme under grant agreement 653965 (AARC)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Contention-aware application performance prediction for disaggregated memory systems

  • Vieira Zacarias, Felippe
  • Nishtala, Rajiv
  • Carpenter, Paul Matthew
Disaggregated memory has recently been proposed as a way to allow flexible and fine-grained allocation of memory capacity to compute jobs. This paper makes an important step towards effective resource allocation on disaggregated memory systems. Specifically, we propose a generic approach to predict the performance degradation due to sharing of disaggregated memory. In contrast to prior work, cache capacity is not shared among multiple applications, which removes a major contributor to application performance. For this reason, our analysis is driven by the demand for memory bandwidth, which has been shown to have an important effect on application performance. We show that profiling the application slowdown often involves significant experimental error and noise, and to this end, we improve the accuracy by linear smoothing of the sensitivity curves. We also show that contention is sensitive to the ratio between read and write memory accesses, and we address this sensitivity by building a family of sensitivity curves according to the read/write ratios. Our results show that the methodology predicts the slowdown in application performance subject to memory contention with an average error of 1.19% and max error of 14.6%. Compared with state-of-the-art, the relative improvements are almost 24% on average and 33% for the worst case., This work is part of a project that has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 754337 (EuroEXA); it has been supported by the Spanish Ministry of Science and Innovation (project TIN2015-65316-P and Ramon y Cajal fellowship RYC2018-025628-I), Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR1272), and the Severo Ochoa Programme (SEV-2015-0493)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Performance characterization of spark workloads on shared NUMA Systems

  • Baig, Shuja-ur-Rehman
  • Amaral, Marcelo
  • Polo Cantero, José
  • Carrera Pérez, David
As the adoption of Big Data technologies becomes the norm in an increasing number of scenarios, there is also a growing need to optimize them for modern processors. Spark has gained momentum over the last few years among companies looking for high performance solutions that can scale out across different cluster sizes. At the same time, modern processors can be connected to large amounts of physical memory, in the range of up to few terabytes. This opens an enormous range of opportunities for runtimes and applications that aim to improve their performance by leveraging low latencies and high bandwidth provided by RAM. The result is that there are several examples today of applications that have started pushing the in-memory computing paradigm to accelerate tasks. To deliver such a large physical memory capacity, hardware vendors have leveraged Non-Uniform Memory Architectures (NUMA). This paper explores how Spark-based workloads are impacted by the effects of NUMA-placement decisions, how different Spark configurations result in changes in delivered performance, how the characteristics of the applications can be used to predict workload collocation conflicts, and how to improve performance by collocating workloads in scale-up nodes. We explore several workloads run on top of the IBM Power8 processor, and provide manual strategies that can leverage performance improvements up to 40% on Spark workloads when using smart processor-pinning and workload collocation strategies., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Parallelware tools: an experimental evaluation on POWER systems

  • Arenaz Silva, Manuel
  • Martorell Bofill, Xavier
Static code analysis tools are designed to aid software developers to build better quality software in less time, by detecting defects early in the software development life cycle. Even the most experienced developer regularly introduces coding defects. Identifying, mitigating and resolving defects is an essential part of the software development process, but frequently defects can go undetected. One defect can lead to a minor malfunction or cause serious security and safety issues. This is magnified in the development of the complex parallel software required to exploit modern heterogeneous multicore hardware. Thus, there is an urgent need for new static code analysis tools to help in building better concurrent and parallel software. The paper reports preliminary results about the use of Appentra’s Parallelware technology to address this problem from the following three perspectives: finding concurrency issues in the code, discovering new opportunities for parallelization in the code, and generating parallel-equivalent codes that enable tasks to run faster. The paper also presents experimental results using well-known scientific codes and POWER systems., This work has been partly funded from the Spanish Ministry of Science and Technology (TIN2015-65316-P), the Departament d’Innovació, Universitats i Empresa de la Generalitat de Catalunya (MPEXPAR: Models de Programació i Entorns d’Execució Parallels, 2014-SGR-1051), and the European Union’s Horizon 2020 research and innovation program throughgrant agreements MAESTRO (801101) and EPEEC (801051)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Adaptive sliding windows for improved estimation of data center resource utilization

  • Baig, Shuja-ur-Rehman
  • Iqbal, Waheed
  • Berral García, Josep Lluís
  • Carrera Pérez, David
Accurate prediction of data center resource utilization is required for capacity planning, job scheduling, energy saving, workload placement, and load balancing to utilize the resources efficiently. However, accurately predicting those resources is challenging due to dynamic workloads, heterogeneous infrastructures, and multi-tenant co-hosted applications. Existing prediction methods use fixed size observation windows which cannot produce accurate results because of not being adaptively adjusted to capture local trends in the most recent data. Therefore, those methods train on large fixed sliding windows using an irrelevant large number of observations yielding to inaccurate estimations or fall for inaccuracy due to degradation of estimations with short windows on quick changing trends. In this paper we propose a deep learning-based adaptive window size selection method, dynamically limiting the sliding window size to capture the trend for the latest resource utilization, then build an estimation model for each trend period. We evaluate the proposed method against multiple baseline and state-of-the-art methods, using real data-center workload data sets. The experimental evaluation shows that the proposed solution outperforms those state-of-the-art approaches and yields 16 to 54% improved prediction accuracy compared to the baseline methods., This work is partially supported by the European ResearchCouncil (ERC) under the EU Horizon 2020 programme(GA 639595), the Spanish Ministry of Economy, Industry andCompetitiveness (TIN2015-65316-P and IJCI2016-27485), theGeneralitat de Catalunya, Spain (2014-SGR-1051) and Universityof the Punjab, Pakistan. The statements made herein are solelythe responsibility of the authors., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Deep learning is blowing in the wind. Deep models applied to wind prediction at turbine level

  • Manero Font, Jaume
  • Béjar Alonso, Javier
  • Cortés García, Claudio Ulises
Wind Energy generation depends on the existence of wind, a meteorological phenomena intermittent by nature, with the consequence of generating uncertainty on the availability of wind energy in the future. The grid stability processes require continuous forecasting of wind energy generated. Forecasting wind energy can be performed either by using weather forecast data or by projecting (or regressing) the past time-series data observations into the future. This last method is the statistical or time series approach. Wind Time Series show non-linearity and non-stationarity properties, and these two properties increase the complexity of the forecasting task using statistical methodologies. In this paper we explore the use of deep learning techniques, which can represent non-linearity, to the wind speed prediction using the largest public wind dataset available, the Wind Toolkit from the National Renewable Laboratory of the US. Several deep network architectures like Multi Layer Perceptrons, Convolutional Networks or Recurrent Networks have been tested on the 126,692 wind-sites and with the results obtained valuable comparisons and conclusions have been obtained. The distribution of the wind sites across the North American Geography has allowed to include in the analysis relationships between terrain, wind forecast complexity and deep methods. With the developed testing workbench and with the availability of the Barcelona Supercomputing Center new architectures are being developed. This work concludes with the feasibility of deep learning architectures for the wind and energy forecasting., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

A programming model for hybrid workflows: combining task-based workflows and dataflows all-in-one

  • Ramón Cortés, Cristian
  • Lordan Gomis, Francesc
  • Ejarque Artigas, Jorge
  • Badia Sala, Rosa Maria
In the past years, e-Science applications have evolved from large-scale simulations executed in a single cluster to more complex workflows where these simulations are combined with High-Performance Data Analytics (HPDA). To implement these workflows, developers are currently using different patterns; mainly task-based and dataflow. However, since these patterns are usually managed by separated frameworks, the implementation of these applications requires to combine them; considerably increasing the effort for learning, deploying, and integrating applications in the different frameworks. This paper tries to reduce this effort by proposing a way to extend task-based management systems to support continuous input and output data to enable the combination of task-based workflows and dataflows (Hybrid Workflows from now on) using a single programming model. Hence, developers can build complex Data Science workflows with different approaches depending on the requirements. To illustrate the capabilities of Hybrid Workflows, we have built a Distributed Stream Library and a fully functional prototype extending COMPSs, a mature, general-purpose, task-based, parallel programming model. The library can be easily integrated with existing task-based frameworks to provide support for dataflows. Also, it provides a homogeneous, generic, and simple representation of object and file streams in both Java and Python; enabling complex workflows to handle any data type without dealing directly with the streaming back-end. During the evaluation, we introduce four use cases to illustrate the new capabilities of Hybrid Workflows; measuring the performance benefits when processing data continuously as it is generated, when removing synchronisation points, when processing external real-time data, and when combining task-based workflows and dataflows at different levels. The users identifying these patterns in their workflows may use the presented uses cases (and their performance improvements) as a reference to update their code and benefit of the capabilities of Hybrid Workflows. Furthermore, we analyse the scalability in terms of the number of writers and readers and measure the task analysis, task scheduling, and task execution times when using objects or streams., This work has been supported by the Spanish Government (contracts SEV2015-0493 and TIN2015-65316-P), by Generalitat de Catalunya (contract 2014-SGR-1051), and by the European Commission through the Horizon 2020 Research and Innovation program under contract 730929 (MF2C project). Cristian Ramon-Cortes predoctoral contract is financed by the Spanish Government under the contract BES-2016-076791., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

On the maturity of parallel applications for asymmetric multi-core processors

  • Chronaki, Kallia
  • Moretó Planas, Miquel
  • Casas, Marc
  • Rico, Alejandro
  • Badia Sala, Rosa Maria
  • Ayguadé Parra, Eduard
  • Valero Cortés, Mateo
Asymmetric multi-cores (AMCs) are a successful architectural solution for both mobile devices and supercomputers. By maintaining two types of cores (fast and slow) AMCs are able to provide high performance under the facility power budget. This paper performs the first extensive evaluation of how portable are the current HPC applications for such supercomputing systems. Specifically we evaluate several execution models on an ARM big.LITTLE AMC using the PARSEC benchmark suite that includes representative highly parallel applications. We compare schedulers at the user, OS and runtime levels, using both static and dynamic options and multiple configurations, and assess the impact of these options on the well-known problem of balancing the load across AMCs. Our results demonstrate that scheduling is more effective when it takes place in the runtime system level as it improves the baseline by 23%, while the heterogeneous-aware OS scheduling solution improves the baseline by 10%., This work has been supported by the RoMoL ERC Advanced Grant (GA 321253), by the European HiPEAC Network of Excellence, by the Spanish Ministry of Science and Innovation (contracts TIN2015-65316-P), by the Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272), and by the European Union's Horizon 2020 research and innovation programme under grant agreement No 671697 and No. 779877. M. Moretó has been partially supported by the Ministry of Economy and Competitiveness under Ramon y Cajal fellowship number RYC-2016-21104., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

A parallel dynamic asynchronous framework for uncertainty quantification by hierarchical Monte Carlo algorithms

  • Tosi, Riccardo
  • Amela Milian, Ramon
  • Badia Sala, Rosa Maria
  • Rossi, Riccardo
The necessity of dealing with uncertainties is growing in many different fields of science and engineering. Due to the constant development of computational capabilities, current solvers must satisfy both statistical accuracy and computational efficiency. The aim of this work is to introduce an asynchronous framework for Monte Carlo and Multilevel Monte Carlo methods to achieve such a result. The proposed approach presents the same reliability of state of the art techniques, and aims at improving the computational efficiency by adding a new level of parallelism with respect to existing algorithms: between batches, where each batch owns its hierarchy and is independent from the others. Two different numerical problems are considered and solved in a supercomputer to show the behavior of the proposed approach., This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 800898. This work has been supported by the Spanish Government (contracts SEV2015- 0493 and TIN2015-65316-P), by the Generalitat de Catalunya (contract 2014-SGR-1051). The authors thankfully acknowledge the computer resources at MareNostrum and the technical support provided by Barcelona Supercomputing Center (IM-2020-1-0016)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

BLAS-3 optimized by OmpSs regions (LASs library)

  • Valero Lara, Pedro
  • Catalán Pallarés, Sandra
  • Martorell Bofill, Xavier
  • Labarta Mancho, Jesús José
In this paper we propose a set of optimizations for the BLAS-3 routines of LASs library (Linear Algebra routines on OmpSs) and perform a detailed analysis of the impact of the proposed changes in terms of performance and execution time. OmpSs allows to use regions in the dependences of the tasks. This helps not only in the programming of the algorithmic optimizations, but also in the reduction of the execution time achieved by such optimizations. Different strategies are implemented in order to reduce the amount of tasks created (when there is enough parallelism) during the execution of BLAS-3 operations in the original LASs. Also a better IPC is obtained thanks to a better memory hierarchy exploitation. More specifically, we increase the performance, in particular on big matrices, about 12% for TRSM, and 17% for GEMM with respect to the original version of LASs, even using less cores in the case of GEMM/SYMM. Moreover, when LASs is compared to the OpenMP reference dense linear algebra library PLASMA, performance is increased up to 12.5% for GEMM/SYMM, while for TRSM/TRMM this value raises to 15%., This project has received funding from the Spanish Ministry of Economy and Competitiveness under the project Computacion de Altas Prestaciones VII (TIN2015-65316-P), the Departament d’Innovacio, Universitats i Empresa de la Generalitat de Catalunya, under project MPEXPAR: Models de Programacio i Entorns d’Execució Parallels (2014-SGR-1051), and the Juan de la Cierva Grant Agreement No IJCI2017-33511. We also acknowledge the funding provided by Fujitsu under the BSC-Fujitsu joint project: Math Libraries Migration and Optimization., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Variable batched DGEMM

  • Valero-Lara, Pedro
  • Martinez-Perez, Ivan
  • Mateo, Sergio
  • Sirvent Pardell, Raül
  • Beltran Querol, Vicenç
  • Martorell Bofill, Xavier
  • Labarta Mancho, Jesús José
Many scientific applications are in need to solve a high number of small-size independent problems. These individual problems do not provide enough parallelism and then, these must be computed as a batch. Today, vendors such as Intel and NVIDIA are developing their own suite of batch routines. Although most of the works focus on computing batches of fixed size, in real applications we can not assume a uniform size for all set of problems. We explore and analyze different strategies based on parallel for, task and taskloop OpenMP pragmas. Although these strategies are straightforward from a programmer's point of view, they have a different impact on performance. We also analyze a new prototype provided by Intel (MKL), which deals with batch operations (cblas dgemm batch). We propose a new approach called grouping. It basically groups a set of problems until filling a limit in terms of memory occupancy or number of operations. In this way, groups composed by different number of problems are distributed on cores, achieving a more balanced distribution in terms of computational cost. This strategy is able to be up to 6× faster than the Intel (MKL) batch routine, This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 720270 (HBP SGA1), from the Spanish Ministry of Economy and Competitiveness under the project Computacion de Altas Prestaciones VII (TIN2015-65316-P) and the Departament d’Innovacio, Universitats i Empresa de la Generalitat de Catalunya, under project MPEXPAR: Models de Programacio i Entorns d’Execucio Paral·lels (2014-SGR-1051)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

AutoParallel: Automatic parallelisation and distributed execution of affine loop nests in Python

  • Ramón Cortés, Cristian
  • Amela Milian, Ramon
  • Ejarque Artigas, Jorge
  • Clauss, Philippe
  • Badia Sala, Rosa Maria
The last improvements in programming languages and models have focused on simplicity and abstraction; leading Python to the top of the list of the programming languages. However, there is still room for improvement when preventing users from dealing directly with distributed and parallel computing issues. This paper proposes and evaluates AutoParallel, a Python module to automatically find an appropriate task-based parallelisation of affine loop nests and execute them in parallel in a distributed computing infrastructure. It is based on sequential programming and contains one single annotation (in the form of a Python decorator) so that anyone with intermediate-level programming skills can scale up an application to hundreds of cores. The evaluation demonstrates that AutoParallel goes one step further in easing the development of distributed applications. On the one hand, the programmability evaluation highlights the benefits of using a single Python decorator instead of manually annotating each task and its parameters or, even worse, having to develop the parallel code explicitly (e.g., using OpenMP, MPI). On the other hand, the performance evaluation demonstrates that AutoParallel is capable of automatically generating task-based workflows from sequential Python code while achieving the same performances than manually taskified versions of established state-of-the-art algorithms (i.e., Cholesky, LU, and QR decompositions). Finally, AutoParallel is also capable of automatically building data blocks to increase the tasks’ granularity; freeing the user from creating the data chunks, and re-designing the algorithm. For advanced users, we believe that this feature can be useful as a baseline to design blocked algorithms., This work has been supported by the Spanish Government through contracts SEV2015-0493 and TIN2015-65316-P, and by Generalitat de Catalunya through contract 2014-SGR-1051. Cristian Ramon-Cortes predoctoral contract is financed by the Ministry of Economy and Competitiveness under the contract BES-2016-076791., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication

A methodology approach to compare performance of parallel programming models for shared-memory architectures

  • Utrera Iglesias, Gladys Miriam
  • Gil, Marisa
  • Martorell Bofill, Xavier
The majority of current HPC applications are composed of complex and irregular data structures that involve techniques such as linear algebra, graph algorithms, and resource management, for which new platforms with varying computation-unit capacity and features are required. Platforms using several cores with different performance characteristics make a challenge the selection of the best programming model, based on the corresponding executing algorithm. To make this study, there are approaches in the literature, that go from comparing in isolation the corresponding programming models’ primitives to the evaluation of a complete set of benchmarks. Our study shows that none of them may provide enough information for a HPC application to make a programming model selection. In addition, modern platforms are modifying the memory hierarchy, evolving to larger shared and private caches or NUMA regions making the memory wall an issue to consider depending on the memory access patterns of applications. In this work, we propose a methodology based on Parallel Programming Patterns to consider intra and inter socket communication. In this sense, we analyze MPI, OpenMP and the hybrid solution MPI/OpenMP in shared-memory environments. We demonstrate that the proposed comparison methodology may give more accurate predictions in performance for given HPC applications and consequently a useful tool to select the appropriate parallel programming model., This research was supported by the following grants Spanish Ministry of Science and Innovation (contract TIN2015-65316), the Generalitat de Catalunya (2014- SGR-1051) and the the European Commission through the HiPEAC-3 Network of Excellence (FP7/ICT-217068)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Efficient development of high performance data analytics in Python

  • Álvarez Cid-Fuentes, Javier
  • Alvarez, Pol
  • Amela Milian, Ramon
  • Ishii, Kuninori
  • Morizawa, Rafael K.
  • Badia Sala, Rosa Maria
Our society is generating an increasing amount of data at an unprecedented scale, variety, and speed. This also applies to numerous research areas, such as genomics, high energy physics, and astronomy, for which large-scale data processing has become crucial. However, there is still a gap between the traditional scientific computing ecosystem and big data analytics tools and frameworks. On the one hand, high performance computing (HPC) programming models lack productivity, and do not provide means for processing large amounts of data in a simple manner. On the other hand, existing big data processing tools have performance issues in HPC environments, and are not general-purpose. In this paper, we propose and evaluate PyCOMPSs, a task-based programming model for Python, as an excellent solution for distributed big data processing in HPC infrastructures. Among other useful features, PyCOMPSs offers a highly productive general-purpose programming model, is infrastructure-agnostic, and provides transparent data management with support for distributed storage systems. We show how two machine learning algorithms (Cascade SVM and K-means) can be developed with PyCOMPSs, and evaluate PyCOMPSs’ productivity based on these algorithms. Additionally, we evaluate PyCOMPSs performance on an HPC cluster using up to 1,536 cores and 320 million input vectors. Our results show that PyCOMPSs achieves similar performance and scalability to MPI in HPC infrastructures, while providing a much more productive interface that allows the easy development of data analytics algorithms., This work has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement H2020-MSCA-COFUND2016-754433. This work has been supported by the Spanish Government (SEV2015-0493), by the Spanish Ministry of Science and Innovation (contract TIN2015-65316-P), by Generalitat de Catalunya, Spain (contract 2014-SGR-1051). The research leading to these results has also received funding from the collaboration between Fujitsu and BSC (Script Language Platform)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Mask-guided sample selection for semi-supervised instance segmentation

  • Bellver Bueno, Míriam
  • Salvador Aguilera, Amaia
  • Torres Viñals, Jordi
  • Giró Nieto, Xavier
The final publication is available at Springer via http://dx.doi.org/10.1007/s11042-020-09235-4, Image segmentation methods are usually trained with pixel-level annotations, which require significant human effort to collect. Weakly-supervised pipelines are the most common solution to address this constraint because they are trained with lower forms of supervision, such as bounding boxes or scribbles. Semi-supervised methods are another option, that leverage a large amount of unlabeled data and a limited number of strongly-labeled samples. In this second setup, samples to be strongly-annotated can be selected randomly or with an active learning mechanism that chooses the ones that will maximize the model performance. In this work, we propose a sample selection approach to decide which samples to annotate for semi-supervised instance segmentation. Our method consists in first predicting pseudo-masks for the unlabeled pool of samples, together with a score predicting the quality of each mask. This score is an estimate of the Intersection Over Union (IoU) of the segment with the ground truth mask. We study which samples should be annotated based on the quality score, leading to an improved performance for semi-supervised instance segmentation with low annotation budgets., This work was partially supported by the Spanish Ministry of Economy and Competitivity under contracts TIN2012-34557 by the BSC-CNS Severo Ochoa program (SEV-2011-00067), and contracts TEC2013-43935-R and TEC2016-75976-R. It has also been supported by grants 2014-SGR-1051 and 2014-SGR-1421 by the Government of Catalonia, and the European Regional Development Fund (ERDF). We would also like to acknowledge the valuable discussions with Victor Campos., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Full-network embedding in a multimodal embedding pipeline

  • Vilalta Arias, Armand
  • Garcia Gasulla, Dario
  • Parés Pont, Ferran
  • Moreno Vázquez, Jonatan
  • Ayguadé Parra, Eduard
  • Labarta Mancho, Jesús José
  • Cortés García, Claudio Ulises
  • Suzumura, Toyotaro
The current state-of-the-art for image annotation and image retrieval tasks is obtained through deep neural networks, which combine an image representation and a text representation into a shared embedding space. In this paper we evaluate the impact of using the Full-Network embedding in this setting, replacing the original image representation in a competitive multimodal embedding generation scheme. Unlike the one-layer image embeddings typically used by most approaches, the Full-Network embedding provides a multi-scale representation of images, which results in richer characterizations. To measure the influence of the Full-Network embedding, we evaluate its performance on three different datasets, and compare the results with the original multimodal embedding generation scheme when using a one-layer image embedding, and with the rest of the state-of-the-art. Results for image annotation and image retrieval tasks indicate that the Full-Network embedding is consistently superior to the one-layer embedding. These results motivate the integration of the Full-Network embedding on any multimodal embedding generation scheme, something feasible thanks to the flexibility of the approach, This work is partially supported by the Joint Study Agreement no. W156463 under the IBM/BSC Deep Learning Center agreement, by the Spanish Government through Programa Severo Ochoa (SEV-2015- 0493), by the Spanish Ministry of Science and Technology through TIN2015-65316-P project, by the Generalitat de Catalunya (contracts 2014-SGR-1051), and by the Core Research for Evolutional Science and Technology (CREST) program of Japan Science and Technology Agency (JST)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Freezing Time: a new approach for emulating fast storage devices using VM

  • Bona, Luis C.E.
  • Elias, Alessandro
  • Ziviani, Andre P.
  • Cortés, Toni
  • Nou Castell, Ramon
  • Alves, Marco A.Z.
© 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Recently we are seeing a considerable effort from both academy and industry in proposing new technologies for storage devices. Often these devices are not readily available for evaluation and methods to allow performing their tests just from their performance parameters are an important tool for system administrators. Simulators are a traditional approach for carrying out such evaluations, however, they are more suitable for evaluating the storage device as an isolate component, mostly due to time constraints. In this paper, we propose an approach based on virtual machine technology that is capable of emulate storage devices transparently for the operating system allowing evaluation of simulating devices within a real system using any synthetic or real workload. To emulate devices in real environments it is necessary to use the currently available devices as a storage medium which creates a difficulty when the device to be emulated is faster than this storage medium. To circumvent this limitation we introduce a new technique called Freezing Time, which takes advantage of virtual machine pausing mechanism to manipulate the virtual machine clock and hide the real I/O completion time. Our approach can be implemented just requiring the hypervisor to be modified, providing a high degree of compatibility and flexibility since it is not necessary to modify neither the operating system nor the application. We evaluate our tool under a real system using old magnetic disks to emulate faster storage devices. Experiments using our technique presented an average latency error of 6.08% for read operations and 6.78% for write operations when comparing a real to device., This work was partially supported by the Spanish Ministry of Science and Innovation under the TIN2015–65316 grant, the Generalitat de Catalunya under contract 2014–SGR–1051., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Exploiting parallelism on GPUs and FPGAs with OmpSs

  • Bosch Pons, Jaume
  • Filgueras Izquierdo, Antonio
  • Vidal, Miquel
  • Jiménez González, Daniel
  • Álvarez Martínez, Carlos
  • Martorell, Xavier
This paper presents the OmpSs approach to deal with heterogeneous programming on GPU and FPGA accelerators. The OmpSs programming model is based on the Mercurium compiler and the Nanos++ runtime. Applications are annotated with compiler directives specifying task-based parallelism. The Mercurium compiler transforms the code to exploit the parallelism in the SMP host cores, and also to spawn work on CUDA/OpenCL devices, and FPGA accelerators. For the CUDA/OpenCL devices, the programmer needs only to insert the annotations and provide the kernel function to be compiled by the native CUDA/OpenCL compiler. In the case of the FPGAs, OmpSs uses the High-Level Synthesis tools from FPGA vendors to generate the IP configurations for the FPGA. In this paper we present the performance obtained on the matrix multiply benchmark in the Xilinx Zynq Ultrascale+, as a result of using OmpSs on this benchmark., This work is partially supported by the European Union H2020 program through the AXIOM project (grant ICT-01-2014 GA 645496) and HiPEAC (GA 687698), by the Spanish Government through Programa Severo Ochoa (SEV-2011-0067), by the Spanish Ministerio de Economia y Competitividad under contract Computacion de Altas Prestaciones VII (TIN2015-65316-P), and the Departament d’Innovacio, Universitats i Empresa de la Generalitat de Catalunya, under project MPEXPAR: Models de Programacio i Entorns d’Execucio Paral·lels (2014-SGR-1051)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Worksharing tasks: An efficient way to exploit irregular and fine-grained loop parallelism

  • Maroñas Bravo, Marcos
  • Sala Penadés, Kevin
  • Mateo, Sergi
  • Ayguadé Parra, Eduard
  • Beltran, Vicenç
Shared memory programming models usually provide worksharing and task constructs. The former relies on the efficient fork-join execution model to exploit structured parallelism; while the latter relies on fine-grained synchronization among tasks and a flexible data-flow execution model to exploit dynamic, irregular, and nested parallelism. On applications that show both structured and unstructured parallelism, both worksharing and task constructs can be combined. However, it is difficult to mix both execution models without penalizing the data-flow execution model. Hence, on many applications structured parallelism is also exploited using tasks to leverage the full benefits of a pure data-flow execution model. However, task creation and management might introduce a non-negligible overhead that prevents the efficient exploitation of fine-grained structured parallelism, especially on many-core processors. In this work, we propose worksharing tasks. These are tasks that internally leverage worksharing techniques to exploit fine-grained structured loop-based parallelism. The evaluation shows promising results on several benchmarks and platforms., This work is supported by the Spanish Ministerio de Ciencia, Innovacion y Universidades (TIN2015-65316-P), by the Generalitat de Catalunya (2014-SGR-1051) and by the European Union’s Seventh Framework Programme (FP7/2007-2013) and the H2020 funding framework under grant agreement no. H2020-FETHPC-754304 (DEEP-EST)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

DRAM errors in the field: a statistical approach

  • Živanovič, Darko
  • Esmaili Dokht, Pouya
  • Moré, Sergi
  • Bartolomé, Javier
  • Carpenter, Paul Matthew
  • Radojković, Petar
  • Ayguadé Parra, Eduard
This 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 DRAM errors and it compares the reliability of DIMMs from all three major memory manufacturers, built in three different technologies. Our work has two sets of contributions. First, we illustrate the complexity of in-field DRAM error analysis and demonstrate the limitations of various widely-used methods and metrics. For example, we show that average error rates, errors per MB-hour and mean time between failures can provide volatile and unreliable results even after long periods of error logging, leading to incorrect conclusions about DRAM reliability. Second, we present formal statistical methods that overcome many of the limitations of the current approaches. The methods that we present are simple to understand and implement, reliable and widely accepted in the statistical community. Overall, our study alerts the community about the need to, firstly, question the current practice in quantifying DRAM reliability and, secondly, to select a proper analysis approach for future studies. Our strong recommendations are to focus on metrics with a practical value that could be easily related to system reliability, and to select methods that provide stable results, ideally supported with statistical significance., This work was supported by the Collaboration Agreement between Samsung Electronics Co., Ltd. and BSC, Spanish Government through Severo Ochoa programme (SEV-2015-0493), by the Spanish Ministry of Science and Technology through TIN2015-65316-P project and by the Generalitat de Catalunya (contracts 2014-SGR1051 and 2014-SGR-1272). This work has also received funding from the European Union’s Horizon 2020 research and innovation programme under EuroEXA project (grant agreement No 754337). Darko Zivanovic holds the Severo Ochoa grant (SVP-2014-068501) of the Ministry of Economy and Competitiveness of Spain., Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Improving maritime traffic emission estimations on missing data with CRBMs

  • Gutiérrez Torre, Alberto
  • Berral García, Josep Lluís
  • Buchaca Prats, David
  • Guevara Vilardell, Marc
  • Soret, Albert
  • Carrera Pérez, David
© Elsevier 2020. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/, Maritime traffic emissions are a major concern to governments as they heavily impact the Air Quality in coastal cities. Ships use the Automatic Identification System (AIS) to continuously report position and speed among other features, and therefore this data is suitable to be used to estimate emissions, if it is combined with engine data. However, important ship features are often inaccurate or missing. State-of-the-art complex systems, like CALIOPE at the Barcelona Supercomputing Center, are used to model Air Quality. These systems can benefit from AIS based emission models as they are very precise in positioning the pollution. Unfortunately, these models are sensitive to missing or corrupted data, and therefore they need data curation techniques to significantly improve the estimation accuracy. In this work, we propose a methodology for treating ship data using Conditional Restricted Boltzmann Machines (CRBMs) plus machine learning methods to improve the quality of data passed to emission models that can also be applied to other GPS and time-series problems. Results show that we can improve the default methods proposed to cover missing data. In our results, we observed that using our method the models boosted their accuracy to detect otherwise undetectable emissions. In particular, we used a real data-set of AIS data, provided by the Spanish Port Authority, to estimate that thanks to our method, the model was able to detect 45% of additional emissions, representing 152 tonnes of pollutants per week in Barcelona and propose new features that may enhance emission modeling., This project hasreceived funding from the European Research Council (ERC) under the European Union Horizon 2020 research and innovation programme (grant agreement No 639595). This work is partially supported by the Ministry of Economy, Industry and Competitiveness of Spain under contracts TIN2015-65316-P,2014SGR1051, IJCI2016-27485, and Severo Ochoa Center of Excellence SEV-2015-0493-16-5., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Zero-shot fashion products clustering on social image streams

  • Poveda Pena, Jonatan
  • Tous Liesa, Rubén
Computer Vision methods have been proposed to solve the problem of matching photographs containing some products from users in social media to products in retail catalogues. This is challenging due to the quality of the photographies, difficulties in dealing with garments and their category taxonomy. A N-Shot Learning approach is required as retail catalogues may contain hundreds of different products for which, in many cases, only one image is provided. This framework can be solved by means of Deep Metric Learning (DML) techniques, in which a metric to discriminate similar than dissimilar samples is learnt. The performance of different authors tackling this problem varies a lot but even if they perform reasonably well, the set of elements they need to return in order to include the exact product is large. As after the query there is a person curating the results, it is important to return the smallest set of elements possible, being ideally just to return only one: the related product. This paper proposes to solve the image-to-product image matching problem through a product retrieval system using DML and Zero-short Learning, focusing on garments, and applying some of the last advances on clustering techniques., This work is partially supported by the Spanish Ministry of Economy and Competi-tivity under contract TIN2015-65316-P and by the SGR programme (2014-SGR-1051and 2017-SGR-962) of the Catalan Government., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Multi-tenant Pub/Sub processing for real-time data streams

  • Villalba Navarro, Álvaro
  • Carrera Pérez, David
Devices and sensors generate streams of data across a diversity of locations and protocols. That data usually reaches a central platform that is used to store and process the streams. Processing can be done in real time, with transformations and enrichment happening on-the-fly, but it can also happen after data is stored and organized in repositories. In the former case, stream processing technologies are required to operate on the data; in the latter batch analytics and queries are of common use. This paper introduces a runtime to dynamically construct data stream processing topologies based on user-supplied code. These dynamic topologies are built on-the-fly using a data subscription model defined by the applications that consume data. Each user-defined processing unit is called a Service Object. Every Service Object consumes input data streams and may produce output streams that others can consume. The subscription-based programing model enables multiple users to deploy their own data-processing services. The runtime does the dynamic forwarding of data and execution of Service Objects from different users. Data streams can originate in real-world devices or they can be the outputs of Service Objects. The runtime leverages Apache STORM for parallel data processing, that combined with dynamic user-code injection provides multi-tenant stream processing topologies. In this work we describe the runtime, its features and implementation details, as well as we include a performance evaluation of some of its core components., This work is partially supported by the European Research Council (ERC) un- der the EU Horizon 2020 programme (GA 639595), the Spanish Ministry of Economy, Industry and Competitivity (TIN2015-65316-P) and the Generalitat de Catalunya (2014-SGR-1051)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication

Towards an energy-aware framework for application development and execution in heterogeneous parallel architectures

  • Djemame, Karim
  • Kavanagh, Richard
  • Kelefouras, Vasilios
  • Aguilà, Adrià
  • Ejarque, Jorge
  • Badia Sala, Rosa Maria
  • García-Pérez, David
  • Pezuela, Clara
  • Deprez, Jean-Christophe
  • Guedria, Lofti
  • De Landtsheer, Renaud
  • Georgiou, Yiannis
The Transparent heterogeneous hardware Architecture deployment for eNergy Gain in Operation (TANGO) project’s goal is to characterise factors which affect power consumption in software development and operation for Heterogeneous Parallel Hardware (HPA) environments. Its main contribution is the combination of requirements engineering and design modelling for self-adaptive software systems, with power consumption awareness in relation to these environments. The energy efficiency and application quality factors are integrated into the application lifecycle (design, implementation and operation). To support this, the key novelty of the project is a reference architecture and its implementation. Moreover, a programming model with built-in support for various hardware architectures including heterogeneous clusters, heterogeneous chips and programmable logic devices is provided. This leads to a new cross-layer programming approach for heterogeneous parallel hardware architectures featuring software and hardware modelling. Application power consumption and performance, data location and time-criticality optimization, as well as security and dependability requirements on the target hardware architecture are supported by the architecture., This work has been supported by the European Commission through the Horizon 2020 Research and Innovation program under contract 687584 (TANGO project) by the Spanish Government under contract TIN2015-65316 and grant SEV-2015-0493 (Severo Ochoa Program) and by Generalitat de Catalunya under contracts 2014- SGR-1051 and 2014-SGR-1272., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

An out-of-the-box full-network embedding for convolutional neural networks

  • Garcia Gasulla, Dario
  • Vilalta Arias, Armand
  • Parés, Ferran
  • Ayguadé Parra, Eduard
  • Labarta Mancho, Jesús José
  • Cortés García, Claudio Ulises
  • Suzumura, Toyotaro
© 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Features extracted through transfer learning can be used to exploit deep learning representations in contexts where there are very few training samples, where there are limited computational resources, or when the tuning of hyper-parameters needed for training deep neural networks is unfeasible. In this paper we propose a novel feature extraction embedding called full-network embedding. This embedding is based on two main points. First, the use of all layers in the network, integrating activations from different levels of information and from different types of layers (i.e., convolutional and fully connected). Second, the contextualisation and leverage of information based on a novel three-valued discretisation method. The former provides extra information useful to extend the characterisation of data, while the later reduces noise and regularises the embedding space. Significantly, this also reduces the computational cost of processing the resultant representations. The proposed method is shown to outperform single layer embeddings on several image classification tasks, while also being more robust to the choice of the pre-trained model used as transfer source., This work is partially supported by the Joint Study Agreement no. W156463 under the IBM/BSC Deep Learning Center agreement, by the Spanish Government through Programa Severo Ochoa (SEV-2015-0493), by the Spanish Ministry of Science and Technology through TIN2015-65316-P project and by the Generalitat de Catalunya (contracts 2014-SGR-1051), and by the Core Research for Evolutional Science and Technology (CREST) program of Japan Science and Technology Agency (JST)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Predicting wind energy generation with recurrent neural networks

  • Manero Font, Jaume
  • Béjar Alonso, Javier
  • Cortés García, Claudio Ulises
Decarbonizing the energy supply requires extensive use of renewable generation. Their intermittent nature requires to obtain accurate forecasts of future generation, at short, mid and long term. Wind Energy generation prediction is based on the ability to forecast wind intensity. This problem has been approached using two families of methods one based on weather forecasting input (Numerical Weather Model Prediction) and the other based on past observations (time series forecasting). This work deals with the application of Deep Learning to wind time series. Wind Time series are non-linear and non-stationary, making their forecasting very challenging. Deep neural networks have shown their success recently for problems involving sequences with non-linear behavior. In this work, we perform experiments comparing the capability of different neural network architectures for multi-step forecasting in a 12 h ahead prediction. For the Time Series input we used the US National Renewable Energy Laboratory’s WIND Dataset [3], (the largest available wind and energy dataset with over 120,000 physical wind sites), this dataset is evenly spread across all the North America geography which has allowed us to obtain conclusions on the relationship between physical site complexity and forecast accuracy. In the preliminary results of this work it can be seen a relationship between the error (measured as R2R2 ) and the complexity of the terrain, and a better accuracy score by some Recurrent Neural Network Architectures., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Towards on-demand I/O forwarding in HPC platforms

  • Bez, Jean Luca
  • Zanon Boito, Francieli
  • Miranda Bueno, Alberto
  • Nou Castell, Ramon
  • Cortés, Toni
  • Navaux, Philippe O.A.
I/O forwarding is an established and widely-adopted technique in HPC to reduce contention and improve I/O performance in the access to shared storage infrastructure. On such machines, this layer is often physically deployed on dedicated nodes, and their connection to the clients is static. Furthermore, the increasingly heterogeneous workloads entering HPC installations stress the I/O stack, requiring tuning and reconfiguration based on the applications' characteristics. Nonetheless, it is not always feasible in a production system to explore the potential benefits of this layer under different configurations without impacting clients. In this paper, we investigate the effects of I/O forwarding on performance by considering the application's I/O access patterns and system characteristics. We aim to explore when forwarding is the best choice for an application, how many I/O nodes it would benefit from, and whether not using forwarding at all might be the correct decision. To gather performance metrics, explore, and understand the impact of forwarding I/O requests of different access patterns, we implemented FORGE, a lightweight I/O forwarding layer in user-space. Using FORGE, we evaluated the optimal forwarding configurations for several access patterns on MareNostrum 4 (Spain) and Santos Dumont (Brazil) supercomputers. Our results demonstrate that shifting the focus from a static system-wide deployment to an on-demand reconfigurable I/O forwarding layer dictated by application demands can improve I/O performance on future machines., This study was financed by the Coordenaçao de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. It has also received support from the Conselho Nacional de Desenvolvimento Científico e Tecnologico (CNPq), Brazil; It is also partially supported by the Spanish Ministry of Economy and Competitiveness (MINECO) under grants PID2019-107255GB; and the Generalitat de Catalunya under contract 2014–SGR–1051., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Workflow environments for advanced cyberinfrastructure platforms

  • Badia Sala, Rosa Maria
  • Ejarque Artigas, Jorge
  • Lordan Gomis, Francesc
  • Lezzi, Daniele
  • Conejero Bañón, Javier
  • Álvarez Cid-Fuentes, Javier
  • Becerra Fontal, Yolanda
  • Queralt Calafat, Anna
Progress in science is deeply bound to the effective use of high-performance computing infrastructures and to the efficient extraction of knowledge from vast amounts of data. Such data comes from different sources that follow a cycle composed of pre-processing steps for data curation and preparation for subsequent computing steps, and later analysis and analytics steps applied to the results. However, scientific workflows are currently fragmented in multiple components, with different processes for computing and data management, and with gaps in the viewpoints of the user profiles involved. Our vision is that future workflow environments and tools for the development of scientific workflows should follow a holistic approach, where both data and computing are integrated in a single flow built on simple, high-level interfaces. The topics of research that we propose involve novel ways to express the workflows that integrate the different data and compute processes, dynamic runtimes to support the execution of the workflows in complex and heterogeneous computing infrastructures in an efficient way, both in terms of performance and energy. These infrastructures include highly distributed resources, from sensors and instruments, and devices in the edge, to High-Performance Computing and Cloud computing resources. This paper presents our vision to develop these workflow environments and also the steps we are currently following to achieve it., This work has been supported by the Spanish Government (SEV2015-0493), by the Spanish Ministry of Science and Innovation (contract TIN2015-65316-P), by Generalitat de Catalunya (contract 2014-SGR-1051). Javier Conejero postdoctoral contract is co-financed by the Ministry of Economy and Competitiveness under Juan de la Cierva Formacion´ postdoctoral fellowship number FJCI-2015-24651. This work is supported by the H2020 mF2C project (730929) and the CLASS project (780622). The participation of Rosa M Badia in the BDEC2 meetings is supported by the EXDCI project (800957). The dislib library developments are partially funded under the project agreement between BSC and FUJITSU., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

mF2C: Towards a coordinated management of the IoT-fof-cloud continuum

  • Masip Bruin, Xavier
  • Marín Tordera, Eva
  • Juan Ferrer, Ana
  • Queralt Calafat, Anna
  • Jukan, Admela
  • García Almiñana, Jordi
  • Lezzi, Daniele
  • Jensen, Jens
  • Cordeiro, Cristovao
  • Leckey, Alexander
  • Salis, Antonio
  • Guilhot, Denis
  • Cankar, Matic
Fog computing enables location dependent resource allocation and low latency services, while fostering novel market and business opportunities in the cloud sector. Aligned to this trend, we refer to Fog-tocloud (F2C) computing system as a new pool of resources, set into a layered and hierarchical model, intended to ease the entire fog and cloud resources management and coordination. The H2020 project mF2C aims at designing, developing and testing a first attempt for a real F2C architecture. This document outlines the architecture and main functionalities of the management framework designed in the mF2C project to coordinate the execution of services in the envisioned set of heterogeneous anddistributed resources., Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Considerations in using OpenCL on GPUs and FPGAs for throughput-oriented genomics workloads

  • Cadenelli, Nicola
  • Jaksic, Zoran
  • Polo Bardés, Jordà
  • Carrera Pérez, David
The recent upsurge in the available amount of health data and the advances in next-generation sequencing are setting the ground for the long-awaited precision medicine. To process this deluge of data, bioinformatics workloads are becoming more complex and more computationally demanding. For this reasons they have been extended to support different computing architectures, such as GPUs and FPGAs, to leverage the form of parallelism typical of each of such architectures. The paper describes how a genomic workload such as k-mer frequency counting that takes advantage of a GPU can be offloaded to one or even more FPGAs. Moreover, it performs a comprehensive analysis of the FPGA acceleration comparing its performance to a non-accelerated configuration and when using a GPU. Lastly, the paper focuses on how, when using accelerators with a throughput-oriented workload, one should also take into consideration both kernel execution time and how well each accelerator board overlaps kernels and PCIe transferred. Results show that acceleration with two FPGAs can improve both time- and energy-to-solution for the entire accelerated part by a factor of 1.32x. Per contra, acceleration with one GPU delivers an improvement of 1.77x in time-to-solution but of a lower 1.49x in energy-to-solution due to persistently higher power consumption. The paper also evaluates how future FPGA boards with components (i.e., off-chip memory and PCIe) on par with those of the GPU board could provide an energy-efficient alternative to GPUs., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Runtime-assisted cache coherence deactivation in task parallel programs

  • Caheny, Paul
  • Álvarez Martí, Lluc
  • Valero Cortés, Mateo
  • Moretó Planas, Miquel
  • Casas, Marc
With increasing core counts, the scalability of directory-based cache coherence has become a challenging problem. To reduce the area and power needs of the directory, recent proposals reduce its size by classifying data as private or shared, and disable coherence for private data. However, existing classification methods suffer from inaccuracies and require complex hardware support with limited scalability. This paper proposes a hardware/software co-designed approach: the runtime system identifies data that is guaranteed by the programming model semantics to not require coherence and notifies the microarchitecture. The microarchitecture deactivates coherence for this private data and powers off unused directory capacity. Our proposal reduces directory accesses to just 26% of the baseline system, and supports a 64x smaller directory with only 2.8% performance degradation. By dynamically calibrating the directory size our proposal saves 86% of dynamic energy consumption in the directory without harming performance., This work has been supported by the RoMoL ERC Advanced Grant (GA 321253), by the European HiPEAC Network of Excellence, by the Spanish Ministry of Economy and Competitiveness (contract TIN2015-65316-P), by the Generalitat de Catalunya (contracts 2014-SGR-1051 and 2014-SGR-1272) and by the European Unions Horizon 2020 research and innovation programme (grant agreements 671697 and 779877). M. Moreto has been partially supported by the Spanish Ministry of Economy, Industry and Competitiveness under Ramon y Cajal fellowship number RYC-2016-21104., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Auto-tuned OpenCL kernel co-execution in OmpSs for heterogeneous systems

  • Pérez, Borja
  • Stafford, Esteban
  • Bosque Orero, José Luis
  • Beivide Palacio, Ramon
  • Mateo Bellido, Sergi
  • Teruel García, Xavier
  • Martorell Bofill, Xavier
  • Ayguadé Parra, Eduard
© <2018>. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/, The emergence of heterogeneous systems has been very notable recently. The nodes of the most powerful computers integrate several compute accelerators, like GPUs. Profiting from such node configurations is not a trivial endeavour. OmpSs is a framework for task based parallel applications, that allows the execution of OpenCl kernels on different compute devices. However, it does not support the co-execution of a single kernel on several devices. This paper presents an extension of OmpSs that rises to this challenge, and presents Auto-Tune, a load balancing algorithm that automatically adjusts its internal parameters to suit the hardware capabilities and application behavior. The extension allows programmers to take full advantage of the computing devices with negligible impact on the code. It takes care of two main issues. First, the automatic distribution of datasets and the management of device memory address spaces. Second, the implementation of a set of load balancing algorithms to adapt to the particularities of applications and systems. Experimental results reveal that the co-execution of single kernels on all the devices in the node is beneficial in terms of performance and energy consumption, and that Auto-Tune gives the best overall results., Peer Reviewed, Preprint
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Training deep neural networks with low precision input data: a hurricane prediction case study

  • Kahira, Albert
  • Bautista-Gómez, Leonardo
  • Badia Sala, Rosa Maria
Training deep neural networks requires huge amounts of data. The next generation of intelligent systems will generate and utilise massive amounts of data which will be transferred along machine learning workflows. We study the effect of reducing the precision of this data at early stages of the workflow (i.e. input) on both prediction accuracy and learning behaviour of deep neural networks. We show that high precision data can be transformed to low precision before feeding it to a neural network model with insignificant depreciation in accuracy. As such, a high precision representation of input data is not entirely necessary for some applications. The findings of this study pave way for the application of deep learning in areas where acquiring high precision data is difficult due to both memory and computational power constraints. We further use a hurricane prediction case study where we predict the monthly number of hurricanes on the Atlantic Ocean using deep neural networks. We train a deep neural network model that predicts the number of hurricanes, first, by using high precision input data and then by using low precision data. This leads to only a drop in prediction accuracy of less than 2%., This work is partly supported by the Spanish Government through Programa Severo Ochoa (SEV-2015-0493), by the Spanish Ministry of Science and Technology through TIN2015-65316 project, by the Generalitat de Catalunya under contracts 2014-SGR-1051 and 2014-SGR-1272., Peer Reviewed, Postprint (author's final draft)
Dipòsit Digital de Documents de la UAB
Publication Artículo científico (JournalArticle)

PMut, a web-based tool for the annotation of pathological variants on proteins, 2017 update

  • López-Ferrando, Víctor
  • Gazzo, Andrea
  • de la Cruz, Xavier|||0000-0002-9738-8472
  • Orozco López, Modesto|||0000-0002-8608-3278
  • Gelpí, Josep Ll|||0000-0002-0566-7723
We present here a full update of the PMut predictor, active since 2005 and with a large acceptance in the field of predicting Mendelian pathological mutations. PMut internal engine has been renewed, and converted into a fully featured standalone training and prediction engine that not only powers PMut web portal, but that can generate custom predictors with alternative training sets or validation schemas. PMut Web portal allows the user to perform pathology predictions, to access a complete repository of pre-calculated predictions, and to generate and validate new predictors. The default predictor performs with good quality scores (MCC values of 0.61 on 10-fold cross validation, and 0.42 on a blind test with SwissVar 2016 mutations). The PMut portal is freely accessible at http://mmb.irbbarcelona.org/PMut. A complete help and tutorial is available at http://mmb.irbbarcelona.org/PMut/help.