Post-pass partitioning of signal processing programs

Symmetric multiprocessor systems are increasingly common, not only as high-throughput servers, but as a vehicle for executing a single application in parallel in order to reduce its execution latency. This article presents Pedigree, a compilation tool that employs a new partitioning heuristic based on the program dependence graph (PDG). Pedigree creates overlapping, potentially interdependent threads, each executing on a subset of the SMP processors that matches the thread’s available parallelism. A unified framework is used to build threads from procedures, loop nests, loop iterations, and smaller constructs. Pedigree does not require any parallel language support; it is post-compilation tool that reads in object code. The SDIO Signal and Data Processing Benchmark Suite has been selected as an example of real-time, latency-sensitive code. Its coarse-grained data flow parallelism is naturally exploited by Pedigree to achieve speedups of 1.63×/2.13× (mean/max) and 1.71×/2.41× on two and four processors, respectively. There is roughly a 20% improvement over existing techniques that exploit only data parallelism. By exploiting the unidirectional flow of data for coarse-grained pipelining, the synchronization overhead is typically limited to less than 6% for synchronization latency of 100 cycles, and less than 2% for 10 cycles. This research was supported by ONR contract numbers N00014-91-J-1518 and N00014-96-1-0347. We would like to thank the Pittsburgh Supercomputing Center for use of their Alpha systems.     

Performance of Reuse Partitioning in Cellular Systems with Mobile Users

Reuse partitioning (RP) is a simple technique that can be used to increase the traffic capacity in a cellular system. With RP, a cell is divided into several concentric regions, each associated with a different cluster size. In this article, a traffic model is developed to analyze the impact of mobile users on a two-region RP system using fixed channel assignment. The influence of user speed and cell size on the new call blocking probability, P_b, and the call dropping probability, P_d, is investigated. A simpler model that can be used to estimate P_b and P_d in some cases is described. The effect on system capacity of reserving some channels for handoff calls is also studied. It is found that even though prioritized handoff can reduce P_d, it may also degrade the capacity. The accuracy of the analysis is verified using simulation for three mobility models.     

基于STL转换为X3D格式的研究与实现

通过对STL三角网格模型和X3D语法规范的分析与研究,根据X3D对三角形的相关定义和描述,提出了一种将STL三角网格模型转换为X3D格式的算法。该算法在实现将STL转化为X3D的同时,又完成了对STL文件中冗余数据的处理,使得转换后的X3D文件大小相当于原来STL文件大小的1/7左右,非常有利于三角网格模型在网络上的传输和操作。最后,以Java作为开发系统平台,运用Java3D及其装载器组件实现了转换后的X3D的装载和浏览,从而使STL文件可以在Internet环境下实现产品信息共享和可视化的操作。     

A Combination of Optimal Partitioning and Location Prediction to Assist Paging in Mobile Cellular Networks

The correlated both in space and time user mobility behavior can aid significantly in the localization of a moving terminal in the network coverage area. However, there is always some uncertainty in mobile user position and a network-wide search cannot be done. Therefore, a predictive paging scheme must always be combined with a location update strategy and a compatible algorithm to conduct the search. In this paper, we introduce a new strategy that combines an optimal partitioning of the location area (LA) with a model to predict user movements, based on the additional information of the cell where the last interaction between the network and the terminal took place. We study the performance of the strategy under delay bounds and different mobility and call arrival characteristics. It is found that the new scheme further minimizes signaling costs and enhances previous sequential paging algorithms.     

New Alternatives to the Estimation Problem in Hardware-Software Codesign of Complex Embedded Systems: The H.261 Video Co-dec Case Study

As Codesign problems become larger and more realistic, the required time to estimate their solutions turns into an important bottleneck. This paper presents a new approach to improve the traditional estimation techniques, in order to avoid this drawback. The presented method has been successfully tested on a large experimental benchmark, attaining quality levels close to those provided by the Synopsys Behavioral Compiler. Finally, a case study based on the standard H.261 video co-dec is described, proving the convenience of the technique on real-life situations. The obtained results show a significant improvement in the process time, while keeping the good precision and fidelity levels that the traditional estimation models usually offer.     

An Iterative Algorithm for Hardware-Software Partitioning, Hardware Design Space Exploration and Scheduling

The paper proposes a novel heuristic technique for integrated hardware-software partitioning, hardware design space exploration and scheduling. The technique maps an application specified as a task graph on a heterogeneous architecture with an objective to minimize the latency of the task graph subject to the area constraint on the hardware coprocessor. The technique uses an iterative approach where the partitioner decides the processor mapping and HW design points of some tasks. The scheduler then simultaneously decides the processor mapping, HW design point and schedule time of the remaining tasks. There exists a tight coupling between the two design stages allowing them to produce superior quality designs in fewer iterations. The technique accounts for the time overheads due to inter-processor /intra-processor communication and shared memory access conflicts. It can therefore be used for both communication intensive and computation intensive applications. The technique also considers dynamic reconfiguration capability of the hardware coprocessor. The technique performs tradeoff analysis and maps hardware tasks to mutually exclusive temporal segments if this results in lower latency. The effectiveness of the technique is demonstrated by a case study of the JPEG image compression algorithm, comparison with an optimal ILP based approach and experimentation with synthetic graphs.     

Minimizing Average Schedule Length under Memory Constraints by Optimal Partitioning and Prefetching

Over the last 20 years, the performance gap between CPU and memory has been steadily increasing. As a result, a variety of techniques has been devised to hide that performance gap, from intermediate fast memories (caches) to various prefetching and memory management techniques for manipulating the data present in these caches. In this paper we propose a new memory management technique that takes advantage of access pattern information that is available at compile time by prefetching certain data elements before explicitly being requested by the CPU, as well as maintaining certain data in the local memory over a number of iterations. In order to better take advantage of the locality of reference present in loop structures, our technique also uses a new approach to memory by partitioning it and reducing execution to each partition, so that information is reused at much smaller time intervals than if execution followed the usual pattern. These combined approaches—using a new set of memory instructions as well as partitioning the memory—lead to improvements in total execution time of approximately 25% over existing methods.     

Computation Partitioning in Mobile Cloud Computing： A Survey

Mobile devices are increasingly interacting with clouds,and mobile cloud computing has emerged as a new paradigm.An central topic in mobile cloud computing is computation partitioning,which involves partitioning the execution of applications between the mobile side and cloud side so that execution cost is minimized.This paper discusses computation partitioning in mobile cloud computing.We first present the background and system models of mobile cloud computation partitioning systems.We then describe and compare state-of-the-art mobile computation partitioning in terms of application modeling,profiling,optimization,and implementation.We point out the main research issues and directions and summarize our own works.     

FUZZY PARTITIONING OF FEATURE SPACE FOR PATTERN CLASSIFICATION BASED ON SUPERVISED C1USTERING

The choice of a fuzzy partitioning is crucial to the performance of a fuzzy system based on if-then rules. However, most of the existing methods are complicated or lead ,o too many subspaces, which is unfit for the applications of pattern classification. A simple but effective clustering approach is proposed in this paper, which obtains a set of compact subspaces and is applicable for classification problems with higher dimensional feature. Its effectiveness is demonstrated by the experimental results.     

Altering the ratio of dietary palmitic,stearic, and oleic acids in diets with or without whole cottonseed affects nutrient digestibility,energy partitioning,and production responses of dairy cows

The objective of this study was to evaluate the effects of varying the ratio of dietary palmitic (C16:0), stearic (C18:0), and oleic (cis-9 C18:1) acids in basal diets containing soyhulls or whole cottonseed on nutrient digestibility, energy partitioning, and production response of lactating dairy cows. Twenty-four mid-lactation multiparous Holstein cows were used in a split-plot Latin square design. Cows were allocated to a main plot receiving either a basal diet with soyhulls (SH, n = 12) or a basal diet with whole cottonseed (CS, n = 12) that was fed throughout the experiment. Within each plot a 4 × 4 Latin square arrangement of treatments was used in 4 consecutive 21-d periods. Treatments were (1) control (CON; no supplemental fat), (2) high C16:0 supplement [PA; fatty acid (FA) supplement blend provided ～80% C16:0], (3) C16:0 and C18:0 supplement (PA+SA; FA supplement blend provided ～40% C16:0 + ～40% C18:0), and (4) C16:0 and cis-9 C18:1 supplement (PA+OA; FA supplement blend provided ～45% C16:0 + ～35% cis-9 C18:1). Interactions between basal diets and FA treatments were observed for dry matter intake (DMI) and milk yield. Among the SH diets, PA and PA+SA increased DMI compared with CON and PA+OA treatments, whereas in the CS diets PA+OA decreased DMI compared with CON. The PA, PA+SA, and PA+OA treatments increased milk yield compared with CON in the SH diets. The CS diets increased milk fat yield compared with the SH diets due to the greater yield of de novo and preformed milk FA. The PA treatment increased milk fat yield compared with CON, PA+SA, and PA+OA due to the greater yield of mixed-source (16-carbon) milk FA. The PA treatment increased 3.5% fat-corrected milk compared with CON and tended to increase it compared with PA+SA and PA+OA. The CS diets increased body weight (BW) change compared with the SH diets. Additionally, PA+OA tended to increase BW change compared with CON and PA and increased it in comparison with PA+SA. The PA and PA+OA treatments increased dry matter and neutral detergent fiber digestibility compared with PA+SA and tended to increase them compared with CON. The PA+SA treatment reduced 16-carbon, 18-carbon, and total FA digestibility compared with the other treatments. The CS diets increased energy partitioning toward body reserves compared with the SH diets. The PA treatment increased energy partitioning toward milk compared with CON and PA+OA and tended to increase it compared with PA+SA. In contrast, PA+OA increased energy partitioned to body reserves compared with PA and PA+SA and tended to increase it compared with CON. In conclusion, milk yield responses to different combinations of FA were affected by the addition of whole cottonseed in the diet. Among the combinations of C16:0, C18:0, and cis-9 C18:1 evaluated, fat supplements with more C16:0 increased energy output in milk, whereas fat supplements with more cis-9 C18:1 increased energy storage in BW. The combination of C16:0 and C18:0 reduced nutrient digestibility, which most likely explains the lower performance observed compared with other treatments.