Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 20
Encontrada(s) 1 página(s)

Recercat. Dipósit de la Recerca de Catalunya
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación Contribución de congreso (ConferenceOutput)

HPC benchmarking: scaling right and looking beyond the average

  • Radulović, Milan
  • Asifuzzaman, Kazi
  • Carpenter, Paul Matthew
  • Radojković, Petar
  • Ayguadé Parra, Eduard
The final publication is available at Springer via http://dx.doi.org/10.1007/978-3-319-96983-1_10, 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 main performance bottlenecks. In this paper, we execute seven production HPC applications on a production HPC platform, and analyse the key performance bottlenecks: FLOPS performance and memory bandwidth congestion, and the implications on scaling out. We show that the results depend significantly on the number of execution processes and granularity of measurements. We therefore advocate for guidance in the application suites, on selecting the representative scale of the experiments. Also, we propose that the FLOPS performance and memory bandwidth should be represented in terms of the proportions of time with low, moderate and severe utilization. We show that this gives much more precise and actionable evidence than the average., 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
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación

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
Publicación Contribución de congreso (ConferenceOutput)

Continuous-action reinforcement learning for memory allocation in virtualized servers

  • Garrido Platero, Luis Ángel
  • Nishtala, Rajiv
  • Carpenter, Paul Matthew
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 having a node-wide view. There are many solutions for the dynamic memory allocation problem, some of which use machine learning in some form. This paper introduces CAVMem (Continuous-Action Algorithm for Virtualized Memory Management), a proof-of-concept mechanism for a decentralized dynamic memory allocation solution in virtualized nodes that applies a continuous-action reinforcement learning (RL) algorithm called Deep Deterministic Policy Gradient (DDPG). CAVMem with DDPG is compared with other RL algorithms such as Q-Learning (QL) and Deep Q-Learning (DQL) in an environment that models a virtualized node. In order to obtain linear scaling and be able to dynamically add and remove VMs, CAVMem has one agent per VM connected via a lightweight coordination mechanism. The agents learn how much memory to bid for or return, in a given state, so that each VM obtains a fair level of performance subject to the available memory resources. Our results show that CAVMem with DDPG performs better than QL and a static allocation case, but it is competitive with DQL. However, CAVMem incurs significant less training overheads than DQL, making the continuous-action approach a more cost-effective solution., This research 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) and the European Union’s 7th Framework Programme under grant agreement number 610456 (Euroserver). It also received funding from the Spanish Ministry of Science and Technology (project TIN2015-65316-P), Generalitat de Catalunya (contract 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
Publicación 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
Publicación 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)