{"title":"A Reconfigurable Processor PARS and its Compiler","authors":"K. Tanigawa, Ryuji Hada, T. Hironaka, A. Kojima","doi":"10.1109/IWIA.2007.11","DOIUrl":"https://doi.org/10.1109/IWIA.2007.11","url":null,"abstract":"The topics of this paper are the following: (1) to introduce a reconfigurable architecture for general purpose, called as PARS, and (2) its compiler to generate high quality code in a reasonable compile time. PARS architecture can execute various application programs by avoiding the program size problem with the fold-down mechanism and reconfiguration. The mechanism is to fold down the program into fixed size partitions which are partitioned by a compiler. PARS architecture executes these folded partitions one by one. The execution was performed by dynamic reconfiguration. To support such dynamic reconfiguration efficiently, we introduce conditional branch in PARS architecture. To utilize the dynamic reconfiguration ability, PARS architecture requires a good compiler which realize a effective execution flow and achieve high quality on routing and placement. The compiler is required to consider a trade off between compile time and quality of generated code, for avoiding explosion of the compile time, when algorithms used in VLSI CAD are simply applied. In this paper, we show how PARS architecture executes a program effectively and the method how the compiler generates high quality code in a reasonable compile time.","PeriodicalId":324361,"journal":{"name":"Innovative architecture for future generation high-performance processors and systems (iwia 2007)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128237528","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"An Embedded Processor: Is It Ready for High-Performance Computing?","authors":"F. Arakawa","doi":"10.1109/IWIA.2007.14","DOIUrl":"https://doi.org/10.1109/IWIA.2007.14","url":null,"abstract":"High power-performance ratio is the most important factor in high-performance computing, whose performance is limited by its power budget. A SuperH (SH) embedded processor core, SH-X3, implemented in a 90-nm CMOS process running at 600 MHz achieved 1080 Dhrystone MIPS, 4.2 GFLOPS, and 55 M polygons/s. Its power performance ratio reaches to as high as 3000 MIPS/W. It is, therefore, a candidate processor for high-performance computing. This paper focuses on the SH processors' low power features, floating-point architecture and performance, and applicability to the high- performance computing.","PeriodicalId":324361,"journal":{"name":"Innovative architecture for future generation high-performance processors and systems (iwia 2007)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125198751","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
J. Ayala, A. Apavatjrut, David Atienza Alonso, M. López-Vallejo
{"title":"Exploring Temperature-Aware Design of Memory Architectures in VLIW Systems","authors":"J. Ayala, A. Apavatjrut, David Atienza Alonso, M. López-Vallejo","doi":"10.1109/IWIA.2007.7","DOIUrl":"https://doi.org/10.1109/IWIA.2007.7","url":null,"abstract":"This paper presents a thermal model to analyze the temperature evolution in the shared register files found on VLIW systems. The use of this model allows the analysis of several factors that have an strong impact on the heat transfer: layout topology, placement and memory accesses. Finally, some relevant conclusions are obtained after analyzing the thermal behavior of several multimedia applications.","PeriodicalId":324361,"journal":{"name":"Innovative architecture for future generation high-performance processors and systems (iwia 2007)","volume":"255 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125776547","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Impact of Predictive Switching in 2-D Torus Networks","authors":"T. Yoshinaga, H. Murakami, M. Koibuchi","doi":"10.1109/IWIA.2007.8","DOIUrl":"https://doi.org/10.1109/IWIA.2007.8","url":null,"abstract":"Predictive switching is a technique for reducing message latency in parallel computer networks. It tries to decide traversal paths of messages by utilizing a prediction mechanism so that processing time for message headers can be shortened. A key issue of predictive switching is the overhead of prediction failures. This paper presents simple and efficient treatments of prediction failures. Our proposal includes three schemes. The first scheme is arranging predictive and non-predictive routers in a network to safely detect and discard mis-predicted packets. The second is additional hardware to reduce occurrences of mis-predicted packets. The third scheme is to shorten the mis-predicted packets. We show the impact of predictive switching embodying the three schemes for k-ary 2-cubes (k = 8, 16, 32) with dimension- order routing. Our simulation results demonstrate that we can reduce average message latency by minimizing the prediction-failure overhead. Network saturation throughput is also improved when the predictor's accuracy is high.","PeriodicalId":324361,"journal":{"name":"Innovative architecture for future generation high-performance processors and systems (iwia 2007)","volume":"15 3","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113956956","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
T. Watanabe, T. Fujise, K. Mori, K. Iwasawa, I. Nakata
{"title":"Design Assists for Embedded Systems in the COINS Compiler Infrastructure","authors":"T. Watanabe, T. Fujise, K. Mori, K. Iwasawa, I. Nakata","doi":"10.1109/IWIA.2007.13","DOIUrl":"https://doi.org/10.1109/IWIA.2007.13","url":null,"abstract":"Program design of embedded systems requires special considerations such as compact code size, shorter addressing field, low cost parallelization, tuning to application field, and so on. COINS is a compiler infrastructure that makes compiler development easy by providing a code generator based on target machine description for quick retargeting, and providing modularized analysis/optimization methods that can be called from user supplied compiler modules implementing new compiler features. Recently it has been extended to be able to select compact code size, to generate parallelized object code, to supply profiling information, etc. Such features to support embedded systems are discussed in this paper.","PeriodicalId":324361,"journal":{"name":"Innovative architecture for future generation high-performance processors and systems (iwia 2007)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129402462","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Responsive Link for Distributed Real-Time Processing","authors":"N. Yamasaki","doi":"10.1109/IWIA.2007.17","DOIUrl":"https://doi.org/10.1109/IWIA.2007.17","url":null,"abstract":"In this paper, we design and implement responsive link, which is a real-time communication link, for distributed real-time systems including sensor-actuator networked systems, ubiquitous computing systems, robot systems, and mechatronic systems. In order to realize flexible real-time communications, the responsive link has many unique features including priority-based packet overtaking (the packet with higher priority overtakes other packets at each node.), separation of data transmission (data link) and event transmission (event link), independent routing of the data link and the event link, packet acceleration/deceleration using priority replacement (packet priority can be replaced with a new priority level at each node to accelerate/decelerate packets under distributed control), prioritized routing (when multiple packets with different priority levels are sent to the same destination, the different route can be set to realize exclusive communication lines or detours.), dynamically variable link speed (800, 400, 200, 100, 50, 25, 12.5 Mbaud), hot-plug&play, topology transparent, etc. The responsive link is implemented on the responsive processor and a few kinds of RMT processors for distributed real-time systems. The responsive link is now under standardization at ISO/IEC JTCI SC25.","PeriodicalId":324361,"journal":{"name":"Innovative architecture for future generation high-performance processors and systems (iwia 2007)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133434849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Improving Search Speed on Pointer-Based Large Data Structures Using a Hierarchical Clustering Copying Algorithm","authors":"M. Yasugi, T. Yuasa","doi":"10.1109/IWIA.2007.18","DOIUrl":"https://doi.org/10.1109/IWIA.2007.18","url":null,"abstract":"The increasing processor-memory performance gap makes improving the cache locality as important as the virtual memory locality. In many applications, especially in search algorithms on pointer-based large data structures, breadth-first copying algorithms increase cache misses, page faults and TLB misses. Since the depth-first copying only achieves limited locality improvement, several clustering copying algorithms have been proposed. In this paper, we propose \"hierarchical clustering\" which groups data objects at multiple hierarchical levels and provides better locality at both the cache and virtual memory levels of the memory hierarchy. We also propose a new copying algorithm for the hierarchical clustering, which uses multiple scan pointers and bounded workspace. Our copying algorithm almost always outperforms other algorithms; after copying, two representative microbenchmarks, employing a search tree and an array of associative lists, run approximately two (sometimes five) times faster than breadth-first copying.","PeriodicalId":324361,"journal":{"name":"Innovative architecture for future generation high-performance processors and systems (iwia 2007)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131948951","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
T. Shimada, T. Tabata, T. Kitamura, T. Kojima, Y. Nakashima, K. Kise
{"title":"Outline of OROCHI: A Multiple Instruction Set Executable SMT Processor","authors":"T. Shimada, T. Tabata, T. Kitamura, T. Kojima, Y. Nakashima, K. Kise","doi":"10.1109/IWIA.2007.15","DOIUrl":"https://doi.org/10.1109/IWIA.2007.15","url":null,"abstract":"In recent years, enjoying multimedia contents with portable devices become popular. These multimedia processing workloads are too heavy workload for a conventional processor so that current portable devices implement additional dedicated processor for multimedia processing. But we have to left conventional processor to execute OS and miscellaneous processing so that this solution enlarges cost, footprint, and power consumption compared to one chip solution. In this paper, we propose the processor called OROCHI which can execute two instruction sets simultaneously. The processor can execute VLIW instruction set for multimedia processing and conventional instruction set for OS and miscellaneous processing. In this paper, we introduce OROCHI processor which is based on a VLIW pipeline. The processor decodes either of the two instruction sets in its corresponding front end. After that, the processor decomposes and translates the conventional instructions and insert them into available slots in VLIW instructions. By these means, we can successfully unite the two processors of different purposes into one specific processor. As a result, we can reduce hardware cost, footprint, and power consumption to meet the rising demands of portable media processing market.","PeriodicalId":324361,"journal":{"name":"Innovative architecture for future generation high-performance processors and systems (iwia 2007)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121861148","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A. Felch, J. Nageswaran, Ashok Chandrashekar, J. Furlong, N. Dutt, R. Granger, A. Nicolau, A. Veidenbaum
{"title":"Accelerating Brain Circuit Simulations of Object Recognition with CELL Processors","authors":"A. Felch, J. Nageswaran, Ashok Chandrashekar, J. Furlong, N. Dutt, R. Granger, A. Nicolau, A. Veidenbaum","doi":"10.1109/IWIA.2007.10","DOIUrl":"https://doi.org/10.1109/IWIA.2007.10","url":null,"abstract":"Humans outperform computers on many natural tasks including vision. Given the human ability to recognize objects rapidly and almost effortlessly, it is pragmatically sensible to study and attempt to imitate algorithms used by the brain. Analysis of the anatomical structure and physiological operation of brain circuits has led to derivation of novel algorithms that in initial study have successfully addressed issues of known difficulty in visual processing. These algorithms are slow on uni-processor based systems, thwarting attempts to drive real-time robots for behavioral study, but as might be expected of algorithms designed for highly parallel brain architectures, they are intrinsically parallel and lend themselves to efficient implementation across multiple processors. This paper presents an implementation of such parallel algorithms on a CELL processor and further extends it to a low-cost cluster built using the Sony PlayStation 3 (PS3). The paper describes the modeled brain circuitry, derived algorithms, implementation on the PS3, and initial performance evaluation with respect to both speed and visual object recognition efficacy. The results show that a parallel implementation can achieve a 140x performance improvement on a cluster of 3 PS3s, attaining real-time processing delays. More importantly, we show that the improvements scale linearly, or nearly so in practice. These initial findings, while highly promising in their own right, also provide a new platform to enable extended investigation of large scale brain circuit models. Early prototyping of such large scale models has yielded evidence of their efficacy in recognition of time-varying, partially occluded, scale-invariant objects in arbitrary scenes.","PeriodicalId":324361,"journal":{"name":"Innovative architecture for future generation high-performance processors and systems (iwia 2007)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125917194","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Some Initial Explorations into the Hierarchical Multi-core Chip Design Space for HPC Systems","authors":"P. Kogge","doi":"10.1109/IWIA.2007.9","DOIUrl":"https://doi.org/10.1109/IWIA.2007.9","url":null,"abstract":"Multi-core designs have emerged as the dominant trend for commodity and high performance microprocessor chips, in virtually all market segments. This includes the high performance supercomputing arena. Using a particular HPC system as a baseline, this paper performs some initial explorations of how the constraints of chip technology, system-imposed memory and bandwidth, and application characteristics may govern the performance achievable from future HPC systems.","PeriodicalId":324361,"journal":{"name":"Innovative architecture for future generation high-performance processors and systems (iwia 2007)","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133880548","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}