Browsing by Author "Subramaniam, Balaji"

Now showing 1 - 4 of 4
  • Thumbnail Image
    The Green500 List: Escapades to Exascale
    Scogland, Thomas R. W.; Subramaniam, Balaji; Feng, Wu-chun (Department of Computer Science, Virginia Polytechnic Institute & State University, 2011)
    Energy efficiency is now a top priority. The first four years of the Green500 have seen the importance of en- ergy efficiency in supercomputing grow from an afterthought to the forefront of innovation as we near a point where sys- tems will be forced to stop drawing more power. Even so, the landscape of efficiency in supercomputing continues to shift, with new trends emerging, and unexpected shifts in previous predictions. This paper offers an in-depth analysis of the new and shifting trends in the Green500. In addition, the analysis of- fers early indications of the track we are taking toward exas- cale, and what an exascale machine in 2018 is likely to look like. Lastly, we discuss the new efforts and collaborations toward designing and establishing better metrics, method- ologies and workloads for the measurement and analysis of energy-efficient supercomputing.
  • Thumbnail Image
    Load-Varying LINPACK: A Benchmark for Evaluating Energy Efficiency in High-End Computing
    Subramaniam, Balaji; Feng, Wu-chun (Department of Computer Science, Virginia Polytechnic Institute & State University, 2010-12-01)
    For decades, performance has driven the high-end computing (HEC) community. However, as highlighted in recent exascale studies that chart a path from petascale to exascale computing, power consumption is fast becoming the major design constraint in HEC. Consequently, the HEC community needs to address this issue in future petascale and exascale computing systems. Current scientific benchmarks, such as LINPACK and SPEChpc, only evaluate HEC systems when running at full throttle, i.e., 100% workload, resulting in a focus on performance and ignoring the issues of power and energy consumption. In contrast, efforts like SPECpower evaluate the energy efficiency of a compute server at varying workloads. This is analogous to evaluating the energy efficiency (i.e., fuel efficiency) of an automobile at varying speeds (e.g., miles per gallon highway versus city). SPECpower, however, only evaluates the energy efficiency of a single compute server rather than a HEC system; furthermore, it is based on SPEC's Java Business Benchmarks (SPECjbb) rather than a scientific benchmark. Given the absence of a load-varying scientific benchmark to evaluate the energy efficiency of HEC systems at different workloads, we propose the load-varying LINPACK (LV-LINPACK) benchmark. In this paper, we identify application parameters that affect performance and provide a methodology to vary the workload of LINPACK, thus enabling a more rigorous study of energy efficiency in supercomputers, or more generally, HEC.
  • Thumbnail Image
    Metrics, Models and Methodologies for Energy-Proportional Computing
    Subramaniam, Balaji (Virginia Tech, 2015-08-21)
    Massive data centers housing thousands of computing nodes have become commonplace in enterprise computing, and the power consumption of such data centers is growing at an unprecedented rate. Exacerbating such costs, data centers are often over-provisioned to avoid costly outages associated with the potential overloading of electrical circuitry. However, such over provisioning is often unnecessary since a data center rarely operates at its maximum capacity. It is imperative that we realize effective strategies to control the power consumption of the server and improve the energy efficiency of data centers. Adding to the problem is the inability of the servers to exhibit energy proportionality which diminishes the overall energy efficiency of the data center. Therefore in this dissertation, we investigate whether it is possible to achieve energy proportionality at the server- and cluster-level by efficient power and resource provisioning. Towards this end, we provide a thorough analysis of energy proportionality at the server and cluster-level and provide insight into the power saving opportunity and mechanisms to improve energy proportionality. Specifically, we make the following contribution at the server-level using enterprise-class workloads. We analyze the average power consumption of the full system as well as the subsystems and describe the energy proportionality of these components, characterize the instantaneous power profile of enterprise-class workloads using the on-chip energy meters, design a runtime system based on a load prediction model and an optimization framework to set the appropriate power constraints to meet specific performance targets and then present the effects of our runtime system on energy proportionality, average power, performance and instantaneous power consumption of enterprise applications. We then make the following contributions at the cluster-level. Using data serving, web searching and data caching as our representative workloads, we first analyze the component-level power distribution on a cluster. Second, we characterize how these workloads utilize the cluster. Third, we analyze the potential of power provisioning techniques (i.e., active low-power, turbo and idle low-power modes) to improve the energy proportionality. We then describe the ability of active low-power modes to provide trade-offs in power and latency. Finally, we compare and contrast power provisioning and resource provisioning techniques. This thesis sheds light on mechanisms to tune the power provisioned for a system under strict performance targets and opportunities to improve energy proportionality and instantaneous power consumption via efficient power and resource provisioning at the server- and cluster-level.
  • Thumbnail Image
    Towards Energy-Proportional Computing for Enterprise-Class Server Workloads
    Feng, Wu-chun; Subramaniam, Balaji (Department of Computer Science, Virginia Polytechnic Institute & State University, 2012)
    Massive data centers housing thousands of computing nodes have become commonplace in enterprise computing, and the power consumption of such data centers is growing at an unprecedented rate. Adding to the problem is the inability of the servers to exhibit energy proportionality, i.e., provide energy-ecient execution under all levels of utilization, which diminishes the overall energy eciency of the data center. It is imperative that we realize eective strategies to control the power consumption of the server and improve the energy eciency of data centers. With the advent of Intel Sandy Bridge processors, we have the ability to specify a limit on power consumption during runtime, which creates opportunities to design new power-management techniques for enterprise workloads and make the systems that they run on more energy-proportional. In this paper, we investigate whether it is possible to achieve energy proportionality for an enterprise-class server workload, namely SPECpower ssj2008 benchmark, by using Intel's Running Average Power Limit (RAPL) interfaces. First, we analyze the power consumption and characterize the instantaneous power prole of the SPECpower benchmark at a subsystem-level using the on-chip energy meters exposed via the RAPL interfaces. We then analyze the impact of RAPL power limiting on the performance, per-transaction response time, power consumption, and energy eciency of the benchmark under dierent load levels. Our observations and results shed light on the ecacy of the RAPL interfaces and provide guidance for designing power-management techniques for enterprise-class workloads.