Search

Probabilistic Occupancy Map (POM) is a method to estimate a location of multiple people, given images taken by multiple cameras from different angles set at a head level. It is useful even in the case of significant occlusion and can derive a location of people without an appearance model and prerequisite knowledge about the number of people in a detection space. However, the computation time of POM increases according to the area of detection space and number of cameras, which limits its performance. In this paper, we report performance enhancement of POM by applying OpenMP and GPGPU. As a result of evaluating test videos with different area size and cameras, we confirm that OpenMP implementation marks 1.8 to 2.7 times speedup compared to single CPU core implementation for all videos, whereas GPU implementation provides speedup in case of large grid size by a maximum of 7.6 times.

Multiplayer tracking is a key technology for strategic analysis in team sports. Since existing solutions use high-resolution cameras and are also expensive to install, they are primarily for professionals and there are almost no solutions that consumers can easily obtain. We focused on K-Shortest Paths which performs well with low-end cameras located at a tripod-level height, where occasional detection failure may occur. We implemented a prototype that can generate visual tracking results based on the algorithm, investigated factors that influences tracking performance and computation time using multi-camera video sequences, and estimated computation time for tracking multiplayer in several team sports.

We introduce a memory efficient implementation for the NVM-based Hybrid BFS algorithm that merges redundant data structures to a single graph data structure, while offloading infrequent accessed graph data on NVMs based on the detailed analysis of access patterns, and demonstrate extremely fast BFS execution for large-scale unstructured graphs whose size exceed the capacity of DRAM on the machine. Experimental results of Kronecker graphs compliant to the Graph500 benchmark on a 2-way INTEL Xeon E5-2690 machine with 256 GB of DRAM show that our proposed implementation can achieve 4.14 GTEPS for a SCALE31 graph problem with 231 vertices and 235 edges, whose size is 4 times larger than the size of graphs that the machine can accommodate only using DRAM with only 14.99 % performance degradation. We also show that the power efficiency of our proposed implementation achieves 11.8 MTEPS/W. Based on the implementation, we have achieved the 3rd and 4th position of the Green Graph500 list (2014 June) in the Big Data category.

Splitter-based parallel sorting algorithms are known to be highly efficient for distributed sorting due to their low communication complexity. Although using GPU accelerators could help to reduce the computation cost in general, their effectiveness in distributed sorting algorithms on large-scale heterogeneous GPU-based systems remains unclear. We investigate applicability of using GPU devices to the splitter-based algorithms and extend HykSort, an existing splitter-based algorithm by offloading costly computation phases to GPUs. We also handle GPU memory overflows by introducing an iterative approach which sorts multiple chunks and merges them into one array. We evaluate the performance of our implementation with local sort acceleration on the TSUBAME2.5 supercomputer that comprises over 4000 NVIDIA K20x GPUs. Performance evaluation of weak scaling shows that we achieve 389 times speedup with 0.25TB/s throughput when sorting 4TB 64bit integer on 1024 nodes compared to running on 1 node; on the other hand, for CPU vs. GPU comparison, our implementation achieves only 1.40 times speedup using 1024 nodes. Detailed analysis however reveals that the limitation is almost entirely due to the bottleneck in CPU-GPU host-to-device bandwidth. With orders of magnitude improvements planned for next generation GPUs, the performance boost will be tremendous in accordance with other successful GPU accelerations.

GPUs can accelerate edge scan performance of graph processing applications; however, the capacity of device memory on GPUs limits the size of graph to process. Efficient techniques to handle GPU memory overflows, including overflow detection and performance analysis in large-scale systems, are not well investigated. To address the problem, we propose a MapReduce-based out-of-core GPU memory management technique for processing large-scale graph applications on heterogeneous GPU-based supercomputers. Our proposed technique automatically handles memory overflows from GPUs by dynamically dividing graph data into multiple chunks and overlaps CPU-GPU data transfer and computation on GPUs as much as possible. Our experimental results on TSUBAME2.5 using 1024 nodes (12288 CPU cores, 3072 GPUs) exhibit that our GPU-based implementation performs 2.10x faster than running on CPU when graph data size does not fit on GPUs. We also study the performance characteristics of our proposed out-of-core GPU memory management technique, including application’s performance and power efficiency of scale-up and scale-out approaches in terms of the number of GPUs. Keywords—Large-scale Graph Processing; GPGPU; MapReduce; Out-of-core Algorithms; Big Data Applications

NVM devices will greatly expand the possibility of processing extremely large-scale graphs that exceed the DRAM capacity of the nodes, however, efficient implementation based on detailed performance analysis of access patterns of unstructured graph kernel on systems that utilize a mixture of DRAM and NVM devices has not been well investigated. We introduce a graph data offloading technique using NVMs that augment the hybrid BFS (Breadth-first search) algorithm widely used in the Graph500 benchmark, and conduct performance analysis to demonstrate the utility of NVMs for unstructured data. Experimental results of a Scale27 problem of a Kronecker graph compliant to the Graph500 benchmark show that our approach maximally sustains 4.22 Giga TEPS (Traversed Edges Per Second), reducing DRAM size by half with only 19.18% performance degradation on a 4-way AMD Opteron 6172 machine heavily equipped with NVM devices. Although direct comparison is difficult, this is significantly greater than the result of 0.05 GTEPS for a SCALE 36 problem by using 1TB of DRAM and 12 TB of NVM as reported by Pearce et al. Although our approach uses higher DRAM to NVM ratio, we show that a good compromise is achievable between performance vs. capacity ratio for processing large-scale graphs. This result as well as detailed performance analysis of the proposed technique suggests that we can process extremely large-scale graphs per node with minimum performance degradation by carefully considering the data structures of a given graph and the access patterns to both DRAM and NVM devices. As a result, our implementation has achieved 4.35 MTEPS/W(Mega TEPS per Watt) and ranked 4th on November 2013 edition of the Green Graph500 list in the Big Data category by using only a single fat server heavily equipped with NVMs.

Graph500 is a new benchmark for supercomputers based on large-scale graph analysis, which is becoming an important form of analysis in many real-world applications. Graph algorithms run well on supercomputers with shared memory. For the Linpack-based supercomputer rankings, TOP500 reports that heterogeneous and distributed-memory super-computers with large numbers of GPGPUs are becoming dominant. However, the performance characteristics of large-scale graph analysis benchmarks such as Graph500 on distributed-memory supercomputers have so far received little study. This is the first report of a performance evaluation and analysis for Graph500 on a commodity-processor-based distributed-memory supercomputer. We found that the reference implementation “replicated-csr” based on distributed level-synchronized breadth-first search solves a large free graph problem with 231 vertices and 235 edges (approximately 2.15 billon vertices and 34.3 billion edges) in 3.09 seconds with 128 nodes and 3,072 cores. This equates to 11 giga-edges traversed per second. We describe the algorithms and implementations of the reference implementations of Graph500, and analyze the performance characteristics with varying graph sizes and numbers of computer nodes and different implementations. Our results will also contribute to the development of optimized algorithms for the coming exascale machines.

MapReduce is a programming model that enables efficient massive data processing in large-scale computing environments such as supercomputers and clouds. Such large-scale computers employ GPUs to enjoy its good peak performance and high memory bandwidth. Since the performace of each job is depending on running application characteristics and underlying computing environments, scheduling MapReduce tasks onto CPU cores and GPU devices for efficient execution is difficult. To address this problem, we have proposed a hybrid scheduling technique for GPU-based computer clusters, which minimizes the execution time of a submitted job using dynamic profiles of Map tasks running on CPU cores and GPU devices. We have implemented a prototype of our proposed scheduling technique by extending MapReduce framework, Hadoop. We have conducted some experiments for this prototype by using a K-means application as a benchmark on a supercomputer. The results show that the proposed technique achieves 1.93 times faster than the Hadoop original scheduling algorithm at 64 nodes (1024 CPU cores and 128 GPU devices). The results also indicate that the performance of map tasks, including both CPU and GPU tasks, is significantly affected by the overhead of map task invocation in the Hadoop framework.

Federated storage resources in geographically distributed environments are becoming viable platforms for data-intensive cloud and grid applications. To improveI /O performance in such environments, we propose a novel model-based I/O performance optimization algorithm for data-intensive applications running on a virtual cluster, which determines virtual machine (VM) migration strategies,i.e., when and where a VM should be migrated, while minimizing the expected value of file access time. We solve this problem as a shortest path problem of a weighted direct acyclic graph (DAG), where the weighted vertex represents a location of a VM and expected file access time from the location, and the weighted edge represents a migration of a VM and time. We construct the DAG from our markov model which represents the dependency of files. Our simulation-based studies suggest that our proposed algorithm can achieve higher performance than simple techniques, such as ones that never migrate VMs: 38% or always migrate VMs onto the locations that hold target files: 47%.

Replication in grid file systems can significantly improve I/O performance of data-intensive applications. However, most of existing replication techniques apply to individual files, which may introduce inefficient replication overheads for a large number of files. We propose a file clustering based replication algorithm for grid file systems. Our algorithm groups files according to a relationship of simultaneous accesses between files and stores replicas of the clustered files into storage nodes, to satisfy expected most of future read access times to the clustered files and replication times for individual files being minimized under the given storage capacity limitation. Our experiments on a given grid environment, 20 nodes of 5 sites, suggest that the proposed algorithm achieves accurate file clustering and efficient replica management; our clustering policy with the file cluster size limit of 5120 MB and the storage capacity limit for replicas of 10240 MB exhibits 1.58 times efficiency than the policy that never groups related files. The results also indicate that the overheads required for introducing our algorithm significantly affect I/O performance of running applications.

We propose a model-based optimization algorithm that determines virtual machine(VM) migration strategies, i.e., which VMs should be migrated to which nodes, while minimizing I/O access costs. We solve this problem as a shortest path problem of a direct acyclic graph which minimizes overall data access costs of target file accesses. Our simulation-based studies suggest that the proposed algorithm can achieve higher performance than simple techniques, such as ones that never migrate VMs or always migrate VMs onto the nodes that hold target files.

Replication in grid file systems can significantly improve I/O performance of data-intensive grid applications, but its manual creation and placement would be impractical in a real grid environment involving thousands to millions of files accessed per application. Although automatic determination of where and how many replicas should be created should be decided with regards to application access patterns and network throughputs, thereby achieving high application throughput and minimizing replica space overhead, previous studies have focused on limited parameter spaces in their algorithmic optimizations. We propose an automated replication algorithm that allows most of I/O accesses to be performed within a given time threshold, while simultaneously minimizing the space overhead by replication. Our algorithm models the replication problem as a combinatorial optimization problem, where the constraints are derived from the given access time threshold and various system parameters, while the objective function being to minimize file replication costs. We solve the optimization problem by dynamically monitoring and estimating inter-node link throughputs and file access patterns of running applications. Our simulation-based studies suggest that the proposed algorithm can achieve higher performance than simple techniques, such as ones that always or never create replicas, while keeping storage usage very low. The results also indicate that the proposed automated algorithm can perform comparably with manual replica placement.

Violet InGaN/GaN light emitting diodes (LEDs) was fabricated on semipolar GaN bulk substrates. The output power and external quantum efficiency at a driving current of 20 mA were measured. The first nonpolar m-plane nitride laser diodes (LDs) were realized on low extended defect bulk m-plane GaN substrates.

GNET-1 is a fully programmable network testbed. It provides functions such as wide area network emulation, network instrumentation, traffic shaping, and traffic generation at gigabit Ethernet wire speeds by programming the core FPGA. GNET-1 is a powerful tool for developing network-aware grid software. It is also a network monitoring and traffic-shaping tool that provides high-performance communication over wide area networks. This work describes several sample uses of GNET-1 and presents its architecture.