Now showing items 1-20 of 57

    • A performance perspective on energy efficient HPC links 

      Saravanan, Karthikeyan P.; Carpenter, Paul Matthew; Ramírez Bellido, Alejandro (Association for Computing Machinery (ACM), 2014)
      Conference lecture
      Open Access
      Energy costs are an increasing part of the total cost of ownership of HPC systems. As HPC systems become increasingly energy proportional in an effort to reduce energy costs, interconnect links stand out for their inefficiency. ...
    • A streaming machine description and programming model 

      Carpenter, Paul Matthew; Ródenas Picó, David; Martorell Bofill, Xavier; Ramírez Bellido, Alejandro; Ayguadé Parra, Eduard (2007-07)
      Article
      Restricted access - publisher's policy
      In this paper we present the initial development of a streaming environment based on a programming model and machine description. The stream programming model consists of an extension to the C language and it’s translation ...
    • ACOTES project: Advanced compiler technologies for embedded streaming 

      Duranton, M.; Munk, H.; Ayguadé Parra, Eduard; Bastoul, C.; Carpenter, Paul Matthew; Chamski, Z.; Cohen, A.; Cornero, M.; Dumont, P.; Pop, S.; Pop, A.; Ornstein, A.; Nuzman, D.; Miranda, C.; Martorell Bofill, Xavier; Lindwer, M.; Ladelsky, R.; Ferrer, Roger; Fellahi, M.; Pouchet, L. N; Zaks, A.; Shvadron, U.; Trifunovic, K.; Rohou, E.; Rosen, I.; Ramírez Bellido, Alejandro; Ródenas, D. (2011-04)
      Article
      Open Access
      Streaming applications are built of data-driven, computational components, consuming and producing unbounded data streams. Streaming oriented systems have become dominant in a wide range of domains, including embedded ...
    • Adaptive runtime-assisted block prefetching on chip-multiprocessors 

      García Flores, Víctor; Rico Carro, Alejandro; Villavieja Prados, Carlos; Carpenter, Paul Matthew; Navarro, Nacho; Ramirez, Alex (2016-04-29)
      Article
      Open Access
      Memory stalls are a significant source of performance degradation in modern processors. Data prefetching is a widely adopted and well studied technique used to alleviate this problem. Prefetching can be performed by the ...
    • Aggregating and managing memory capacity across computing nodes in cloud environments 

      Garrido, Luis; Carpenter, Paul Matthew (Barcelona Supercomputing Center, 2017-05-04)
      Conference report
      Open Access
      Managing memory capacity in cloud environments is a challenging issue, mainly due to the temporal variability in virtual machine (VM) memory demand. The Virtual Machine Manager or the hypervisor allocates a portion of the ...
    • Analyzing performance improvements and energy savings in Infiniband architecture using network compression 

      Dickov, Branimir; Pericas, Miquel; Carpenter, Paul Matthew; Navarro, Nacho; Ayguadé Parra, Eduard (Institute of Electrical and Electronics Engineers (IEEE), 2014)
      Conference report
      Restricted access - publisher's policy
      One of the greatest challenges in HPC is total system power and energy consumption. Whereas HPC interconnects have traditionally been designed with a focus on bandwidth and latency, there is an increasing interest in ...
    • Buffer sizing for self-timed stream programs on heterogeneous distributed memory multiprocessors 

      Carpenter, Paul Matthew; Ramírez Bellido, Alejandro; Ayguadé Parra, Eduard (Springer Verlag, 2010)
      Conference report
      Restricted access - publisher's policy
      Stream programming is a promising way to expose concurrency to the compiler. A stream program is built from kernels that communicate only via point-to-point streams. The stream compiler statically allocates these kernels ...
    • Contention-aware application performance prediction for disaggregated memory systems 

      Vieira Zacarias, Felippe; Nishtala, Rajiv; Carpenter, Paul Matthew (Association for Computing Machinery (ACM), 2020)
      Conference report
      Open Access
      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 ...
    • Continuous-action reinforcement learning for memory allocation in virtualized servers 

      Garrido Platero, Luis Ángel; Nishtala, Rajiv; Carpenter, Paul Matthew (Springer, 2019)
      Conference report
      Open Access
      In a virtualized computing server (node) with multiple Virtual Machines (VMs), it is necessary to dynamically allocate memory among the VMs. In many cases, this is done only considering the memory demand of each VM without ...
    • Controlling Network Latency in Mixed Hadoop Clusters: Do We Need Active Queue Management? 

      Fischer e Silva, Renan; Carpenter, Paul Matthew (IEEE, 2016-12-26)
      Conference report
      Open Access
      With the advent of big data, data center applications are processing vast amounts of unstructured and semi-structured data, in parallel on large clusters, across hundreds to thousands of nodes. The highest performance for ...
    • Cost-aware prediction of uncorrected DRAM errors in the field 

      Boixaderas, Isaac; Carpenter, Paul Matthew; Radojković, Petar; Ayguadé Parra, Eduard (Barcelona Supercomputing Center, 2021-05)
      Conference report
      Open Access
      One of the main causes of hardware failure in large-scale clusters is an uncorrected error in main memory [1]–[4]. Node failures are especially problematic in high-performance computing (HPC), where a single tightly-coupled ...
    • 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 Guix, Marc; Carpenter, Paul Matthew; Radojković, Petar; Ayguadé Parra, Eduard (Institute of Electrical and Electronics Engineers (IEEE), 2020)
      Conference report
      Open Access
      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 ...
    • 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 (Association for Computing Machinery (ACM), 2019)
      Conference report
      Open Access
      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 ...
    • Energy Efficient Ethernet on MapReduce Clusters: Packet Coalescing To Improve 10GbE Links 

      Fischer e Silva, Renan; Carpenter, Paul Matthew (IEEE, 2017-10)
      Article
      Open Access
      An important challenge of modern data centers is to reduce energy consumption, of which a substantial proportion is due to the network. Switches and NICs supporting the recent energy efficient Ethernet (EEE) standard are ...
    • Energy optimizing methodologies on heterogeneous data centers 

      Nishtala, Rajiv; Carpenter, Paul Matthew; Petrucci, Vinicius; Martorell Bofill, Xavier (Barcelona Supercomputing Center, 2017-05-04)
      Conference report
      Open Access
      In 2013, U.S. data centers accounted for 2.2% of the country’s total electricity consumption, a figure that is projected to increase rapidly over the next decade. Many important work-loads are interactive, and they demand ...
    • ETP4HPC’s SRA 5 strategic research agenda for High-Performance Computing in Europe 2022: European HPC research priorities 2023-2027 

      Carpenter, Paul Matthew; Casas Guix, Marc; Unsal, Osman Sabri; Radojkovic, Petar; Martorell Bofill, Xavier; Miranda, Alberto; Guitart Fernández, Jordi; Corbalán González, Julita; Peña Monferrer, Antonio José; Bautista Gomez, Leonardo Arturo; Vázquez García, Miguel; Beltran Querol, Vicenç; Queralt Calafat, Anna; Nou Castell, Ramon; Borrell Pol, Ricard; Houzeaux, Guillaume; Serradell Maronda, Kim; Carrera Pérez, David; García Sáez, Artur; Puchol García, Carlos (2022-09)
      Research report
      Open Access
      This document feeds research and development priorities devel-oped by the European HPC ecosystem into EuroHPC’s Research and Innovation Advisory Group with an aim to define the HPC Technology research Work Programme and ...
    • Exploring interconnect energy savings under East-West traffic pattern of MapReduce clusters 

      Fischer e Silva, Renan; Carpenter, Paul Matthew (Institute of Electrical and Electronics Engineers (IEEE), 2015-10-26)
      Conference report
      Open Access
      An important challenge of modern data centers is to reduce energy consumption, of which a substantial proportion is due to the network. Energy Efficient Ethernet (EEE) is a recent standard that aims to ...
    • High Throughput and Low Latency on Hadoop Clusters Using Explicit Congestion Notification: The Untold Truth 

      Fischer e Silva, Renan; Carpenter, Paul Matthew (IEEE, 2017-09-26)
      Conference lecture
      Open Access
      Various extensions of TCP/IP have been proposed to reduce network latency; examples include Explicit Congestion Notification (ECN), Data Center TCP (DCTCP) and several proposals for Active Queue Management (AQM). Combining ...
    • Hipster: hybrid task manager for latency-critical cloud workloads 

      Nishtala, Rajiv; Carpenter, Paul Matthew; Petrucci, Vinicius; Martorell Bofill, Xavier (Institute of Electrical and Electronics Engineers (IEEE), 2017)
      Conference report
      Open Access
      In 2013, U. S. data centers accounted for 2.2% of the country's total electricity consumption, a figure that is projected to increase rapidly over the next decade. Many important workloads are interactive, and they demand ...
    • HPC benchmarking: scaling right and looking beyond the average 

      Radulović, Milan; Asifuzzaman, Kazi; Carpenter, Paul Matthew; Radojković, Petar; Ayguadé Parra, Eduard (Springer, 2018)
      Conference report
      Open Access
      Designing a balanced HPC system requires an understanding of the dominant performance bottlenecks. There is as yet no well established methodology for a unified evaluation of HPC systems and workloads that quantifies the ...