Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 6
Encontrada(s) 1 página(s)

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

Exploiting Reuse Information to Reduce Refresh Energy in On-Chip eDRAM Caches

  • Valero Bresó, Alejandro
  • Duato Marín, José Francisco
  • Sahuquillo Borrás, Julio|||0000-0001-8630-4846
  • Petit Martí, Salvador Vicente|||0000-0003-2426-4134
This work introduces a novel refresh mechanism that leverages reuse information to decide which blocks should be refreshed in an energy-aware eDRAM last-level cache. Experimental results show that, compared to a conventional eDRAM cache, the energy-aware approach achieves refresh energy savings up to 71%, while the reduction on the overall dynamic energy is by 65% with negligible performance losses., This work was supported by the Spanish Ministerio de Economía y Competitividad (MINECO) and Plan E funds, under Grants TIN-2009-14475-C04-01 and TIN2012-38341-C04-01.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Contribución de congreso (ConferenceOutput)

Combining RAM technologies for hard-error recovery in L1 data caches working at very-low power modes

  • Lorente Garcés, Vicente Jesús
  • Sahuquillo Borrás, Julio|||0000-0001-8630-4846
  • Petit Martí, Salvador Vicente|||0000-0003-2426-4134
  • López Rodríguez, Pedro Juan|||0000-0003-4544-955X
  • Valero Bresó, Alejandro
  • Canal, Ramón
  • Duato Marín, José Francisco
Low-power modes in modern microprocessors rely on low frequencies and low voltages to reduce the energy budget. Nevertheless, manufacturing induced parameter variations can make SRAM cells unreliable producing hard errors at supply voltages below Vccmin. Recent proposals provide a rather low fault-coverage due to the fault coverage/overhead trade-off. We propose a new faulttolerant L1 cache, which combines SRAM and eDRAM cells in L1 data caches to provide 100% SRAM hard-error fault coverage. Results show that, compared to a conventional cache and assuming 50% failure probability at low-power mode, leakage and dynamic energy savings are by 85% and 62%, respectively, with a minimal impact on performance., This work was supported by the Spanish MICINN (TIN2010-18368) with the Consolider-Ingenio 2010 Programme co-funded by the European Commission FEDER funds (CSD2006-00046) and co-funded with the Plan E funds (TIN2009-14475-C04-01). Additionaly, it was supported by Generalitat de Catalunya (2009SGR1250), by FP7 program of the European Commission (TRAMS-248789), and by Spanish MINECO (TIN2012-38341-C04-01).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Contribución de congreso (ConferenceOutput)

Understanding cache hierarchy contention in CMPs to improve job scheduling

  • Feliu-Pérez, Josué|||0000-0003-3017-4266
  • Sahuquillo Borrás, Julio|||0000-0001-8630-4846
  • Petit Martí, Salvador Vicente|||0000-0003-2426-4134
  • Duato Marín, José Francisco
In order to improve CMP performance, recent research has focused on scheduling to mitigate contention produced by the limited memory bandwidth. Nowadays, commercial CMPs implement multi-level cache hierarchies where last level caches are shared by at least two cache structures located at the immediately lower cache level. In turn, these caches can be shared by several multithreaded cores. In this microprocessor design, contention points may appear along the whole memory hierarchy. Moreover, this problem is expected to aggravate in future technologies, since the number of cores and hardware threads, and consequently the size of the shared caches increases with each microprocessor generation. In this paper we characterize the impact on performance of the different contention points that appear along the memory subsystem. Then, we propose a generic scheduling strategy for CMPs that takes into account the available bandwidth at each level of the cache hierarchy. The proposed strategy selects the processes to be co-scheduled and allocates them to cores in order to minimize contention effects. The proposal has been implemented and evaluated in a commercial single-threaded quad-core processor with a relatively small two-level cache hierarchy. Despite these potential contention limitations are less than in recent processor designs, compared to the Linux scheduler, the proposal reaches performance improvements up to 9% while these benefits (across the studied benchmark mixes) are always lower than 6% for a memory-aware scheduler that does not take into account the cache hierarchy. Moreover, in some cases the proposal doubles the speedup achieved by the memory-aware scheduler., This work was supported by the Spanish MICINN, Consolider Programme and Plan E funds, as well as European Commission FEDER funds, under Grants CSD2006-00046 and TIN2009-14475-C04-01.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Contribución de congreso (ConferenceOutput)

A New End-to-End Flow-Control Mechanism for High Performance Computing Clusters

  • Prades, Javier|||0000-0003-3349-2200
  • Silla, Federico|||0000-0002-6435-1200
  • Duato, José
  • Fröning, Holger
  • Nuessle, Mondrian
[EN] High Performance Computing usually leverages messaging libraries such as MPI or GASNet in order to exchange data among processes in large-scale clusters. Furthermore, these libraries make use of specialized low-level networking layers in order to retrieve as much performance as possible from hardware interconnects such as InfiniBand or Myrinet, for example. EXTOLL is another emerging technology targeted for high performance clusters. These specialized low-level networking layers require some kind of flow control in order to prevent buffer overflows at the received side. In this paper we present a new flow control mechanism that is able to adapt the buffering resources used by a process according to the parallel application communication pattern and the varying activity among communicating peers. The tests carried out in a 64-node 1024-core EXTOLL cluster show that our new dynamic flow-control mechanism provides extraordinarily high buffer efficiency along with very low overhead, which is reduced between 8 and 10 times., This work was supported by the Spanish MICINN, Plan E funds, under Grant TIN2009-14475-C04-01.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Contribución de congreso (ConferenceOutput)

CU2rCU: towards the Complete rCUDA Remote GPU Virtualization and Sharing Solution

  • Reaño González, Carlos
  • Peña Monferrer, Antonio José
  • Duato Marín, José Francisco
  • Mayo Gual, Rafael
  • Silla, Federico|||0000-0002-6435-1200
  • Quintana-Ortí, Enrique S.|||0000-0002-5454-165X
GPUs are being increasingly embraced by the high performance computing and computational communities as an effective way of considerably reducing execution time by accelerating significant parts of their application codes. However, despite their extraordinary computing capabilities, the adoption of GPUs in current HPC clusters may present certain negative side-effects. In particular, to ease job scheduling in these platforms, a GPU is usually attached to every node of the cluster. In addition to increasing acquisition costs this favors that GPUs may frequently remain idle, as applications usually do not fully utilize them. On the other hand, idle GPUs consume non-negligible amounts of energy, which translates into very poor energy efficiency during idle cycles. rCUDA was recently developed as a software solution to address these concerns. Specifically, it is a middleware that allows transparently sharing a reduced number of GPUs among the nodes in a cluster. rCUDA thus increases the GPU-utilization rate, taking care of job scheduling. While the initial prototype versions of rCUDA demonstrated its functionality, they also revealed several concerns related with usability and performance. With respect to usability, in this paper we present a new component of the rCUDA suite that allows an automatic transformation of any CUDA source code, so that it can be effectively accommodated within this technology. In response to performance, we briefly show some interesting results, which will be deeply analyzed in future publications. The net outcome is a new version of rCUDA that allows, for any CUDA-compatible program, to use remote GPUs in a cluster with minimum overhead., The researchers at UPV were supported by the Spanish MICINN, Plan E funds, under Grant TIN2009-14475-C04-01 and also by PROMETEO from Generalitat Valenciana (GVA) under Grant PROMETEO/2008/060. Researchers at UJI were supported by the Spanish Ministry of Science and FEDER (contract no. TIN2011-23283), and by the Fundacion Caixa- ´ Castello/Bancaixa (no. P1-1B2009-35)
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Contribución de congreso (ConferenceOutput)

PS-Dir: A Scalable Two-Level Directory Cache

  • Valls, Joan J.
  • Ros Bardisa, Alberto
  • Duato Marín, José Francisco
  • Sahuquillo Borrás, Julio|||0000-0001-8630-4846
  • Gómez Requena, María Engracia|||0000-0003-1466-4118
As the number of cores increases in both incoming and future chip multiprocessors, coherence protocols must address novel hardware structures in order to scale in terms of performance, power, and area. It is well known that most blocks accessed by parallel applications are private (i.e., accessed by a single core). These blocks present different directory requirements and behavior than shared blocks. Based on this fact, this paper proposes a two-level directory cache that tracks shared blocks in a small and fast first-level cache and private blocks in a larger and slower second-level cache, namely Shared and Private caches, respectively. Speed and area reasons suggest the use of eDRAM technology much dense but slower than SRAM technology for the Private cache, which in turn brings energy savings. Experimental results for a 16-core system show improvements in performance by 11.1%, in area by 25.4%, and in energy consumption by 20.5% compared to a conventional directory cache., This work has been supported by the Spanish MICINN, as well as European Commission FEDER funds, under Grants CSD2006-00046 and TIN2009-14475-C04