An Energy-Efficient Heterogeneous System for Embedded Learning and Classification

Majumdar, Abhinandan; Cadambi, Srihari; Chakradhar, Srimat

doi:10.1109/les.2010.2100802

Cited by 17 publications

(12 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unlike the clusters of stream processor, each SPE may run a different kernel. Finally, a recent work in [7], integrating an Atom processor coupled to an ION GPU and a FPGA accelerator, studied an energy-efficient system for embedded learning and classification application.…”

Section: Related Work a General Stream Processorsmentioning

confidence: 99%

Low Power and Scalable Many-Core Architecture for Big-Data Stream Computing

Kanoun

Ruggiero

Atienza

et al. 2014

2014 IEEE Computer Society Annual Symposium on VLSI

View full text Add to dashboard Cite

Abstract-In the last years the process of examining large amounts of different types of data, or Big-Data, in an effort to uncover hidden patterns or unknown correlations has become a major need in our society. In this context, stream mining applications are now widely used in several domains such as financial analysis, video annotation, surveillance, medical services, traffic prediction, etc. In order to cope with the Big-Data stream input and its high variability, modern stream mining applications implement systems with heterogeneous classifiers and adapt online to its input data stream characteristics variation. Moreover, unlike existing architectures for video processing and compression applications, where the processing units are reconfigurable in terms of parameters and possibly even functions as the input data is changing, in Big-Data stream mining applications the complete computing pipeline is changing, as entirely new classifiers and processing functions are invoked depending on the input stream. As a result, new approaches of reconfigurable hardware platform architectures are needed to handle Big-Data streams. However, hardware solutions that have been proposed so far for stream mining applications either target high performance computing without any power consideration (i.e., limiting their applicability in small-scale computing infrastructures or current embedded systems), or they are simply dedicated to a specific learning algorithm (i.e., limited to run with a single type of classifiers). Therefore, in this paper we propose a novel low-power manycore architecture for stream mining applications that is able to cope with the dynamic data-driven nature of stream mining applications while consuming limited power. Our exploration indicates that this new proposed architecture is able to adapt to different classifiers complexities thanks to its multiple scalable vector processing units and their re-configurability feature at runtime. Moreover, our platform architecture includes a memory hierarchy optimized for Big-Data streaming and implements modern fine-grained power management techniques over all the different types of cores allowing then minimum energy consumption for each type of executed classifier.

show abstract

Section: Related Work a General Stream Processorsmentioning

confidence: 99%

Low Power and Scalable Many-Core Architecture for Big-Data Stream Computing

Kanoun

Ruggiero

Atienza

et al. 2014

2014 IEEE Computer Society Annual Symposium on VLSI

View full text Add to dashboard Cite

show abstract

“…Good coprocessor utilization, which is the fraction of time the coprocessor is busy, is an important consideration because coprocessors are expensive and power-hungry [17]. We argue that the current GPU usage model does not facilitate achieving high utilization.…”

Section: Introductionmentioning

confidence: 96%

Interference-driven resource management for GPU-based heterogeneous clusters

Phull¹,

Li²,

Rao³

et al. 2012

Proceedings of the 21st International Symposium on High-Performance Parallel and Distributed Computing

View full text Add to dashboard Cite

GPU-based clusters are increasingly being deployed in HPC environments to accelerate a variety of scientific applications. Despite their growing popularity, the GPU devices themselves are under-utilized even for many computationallyintensive jobs. This stems from the fact that the typical GPU usage model is one in which a host processor periodically offloads computationally intensive portions of an application to the coprocessor. Since some portions of code cannot be offloaded to the GPU (for example, code performing network communication in MPI applications), this usage model results in periods of time when the GPU is idle. GPUs could be time-shared across jobs to "fill" these idle periods, but unlike CPU resources such as the cache, the effects of sharing the GPU are not well understood. Specifically, two jobs that time-share a single GPU will experience resource contention and interfere with each other. The resulting slowdown could lead to missed job deadlines. Current cluster managers do not support GPU-sharing, but instead dedicate GPUs to a job for the job's lifetime.In this paper, we present a framework to predict and handle interference when two or more jobs time-share GPUs in HPC clusters. Our framework consists of an analysis model, and a dynamic interference detection and response mechanism to detect excessive interference and restart the interfering jobs on different nodes. We implement our framework in Torque, an open-source cluster manager, and using real workloads on an HPC cluster, show that interference-aware two-job colocation (although our method is applicable to colocating more than two jobs) improves GPU utilization by 25%, reduces a job's waiting time in the queue by 39% and improves job latencies by around 20%.

show abstract

“…Particularly in safety‐critical or medical systems, high reliability is required together with high performance, and reliability is also seriously limited by power consumption . Therefore, there is a strong need for power/energy‐aware design space exploration studies to develop a good understanding of GPU power/energy consumption with the ultimate goal of improving energy efficiency …”

Section: Introductionmentioning

confidence: 99%

“…8,9 Therefore, there is a strong need for power/energy-aware design space exploration studies to develop a good understanding of GPU power/energy consumption with the ultimate goal of improving energy efficiency. [10][11][12][13] The voltage and frequency scaling (VFS) technique is a power-saving method that scales down core frequency or voltage to reduce the power and energy consumption of a digital circuit component without degrading the target application performance or with a tolerable performance degradation. When VFS is performed at runtime dynamically, it is called dynamic VFS (DVFS).…”

Section: Introductionmentioning

confidence: 99%

A performance, power, and energy analysis of ultrasound B‐mode imaging on a GPU with VFS

Phuong

Lee

2016

Concurrency and Computation

View full text Add to dashboard Cite

Summary Because of the power and energy constraints in modern computing systems, it is important to find a balance between performance and power/energy consumption to produce not only energy efficient portable devices but also reliable computing systems. In this paper, we present an analytical investigation of the performance and power/energy consumption of a well‐known medical application, “ultrasound B‐mode imaging,” which is implemented on modern commercial graphics processing units (GPUs). With this target application, we investigate the impact of GPU hardware features on performance, power, and energy by experimentally applying voltage and frequency scaling on various GPU devices. Based on our results, we mathematically build a performance prediction model to show the analytical relationship between the architectural features of those GPUs and the application performance. Our results show that the performance prediction model has errors of less than 8.71% and that there exists an energy‐optimal point in the design space spanned by the core clock frequency. We expect that our model can help engineers to design an ultrasound diagnosis system with the most suitable GPU that satisfies cost, performance, and power/energy constraints.

show abstract

An Energy-Efficient Heterogeneous System for Embedded Learning and Classification

Cited by 17 publications

References 9 publications

Low Power and Scalable Many-Core Architecture for Big-Data Stream Computing

Low Power and Scalable Many-Core Architecture for Big-Data Stream Computing

Interference-driven resource management for GPU-based heterogeneous clusters

A performance, power, and energy analysis of ultrasound B‐mode imaging on a GPU with VFS

Contact Info

Product

Resources

About