CANDAR Keynote 1
- Chair: Yasuhiko Nakashima (NAIST)
- Speaker: Masaya Shida (Socionext Inc.)
- Title: Next Generation Design Framework for Custom Image Processing Systems
- Abstract: We are studying to establish new methods for designing SoC that can meet with each aim of functions and performance for image processing systems we need. For this approach, we are building a new framework for top-down design of image processing. On this framework, we can design specific algorithms for compression/decompression, noise reduction and image correction with C language and can verify them from the view of whole of the system. Also we can evaluate and make analysis both function and performance through carrying out High Level Synthesis and verification on FPGA environment. Then we can make design space exploration for the best combination of SoC(as hardware) and application suitable for specific algorithm. We believe our framework can contribute to configure the best result of function-partitioning and parallelism.
CANDAR Keynote 2
- Chair: Satoshi Fujita (Hiroshima University)
- Speaker: Shin-ichi Minato (Hokkaido University)
- Title: The Art of Graph Algorithms Based on “Power of Enumeration”
- Abstract: Today, in many applications for network analysis and optimization, we often faced with a problem which leads to combinatorial explosion. For demonstrating the power of combinatorial explosion and the art of algorithms to attack the problems, we produced an animation video “Time with class! Let's count!” in YouTube Miraikan Channel. The video got more than 1.6 million views and still now increasing. In this talk, I will present the state-of-the-art techniques of graph enumeration assumed in this video, and also show some prospective applications of the techniques for real-life network analysis.
- Chair: Ferdinand Peper (NICT)
- Speaker: Yuzuru Sato (Hokkaido University)
- Title: Robust computation in neural field
- Abstract: We study robust computation represented by collective motion of large neural networks. There exist stable traveling bumps and their collisions in a two dimensional Amari neural field. By using the stable traveling bumps logical operations and working memory may be constructed. The resulting computation processes is structurally stable, and the basin measure of the dynamics is positive. Thus, these computation are robust and constructive in spatio-temporally continuous activator-inhibitor systems. Experimental data analysis and modeling for membrane potential waves in mouse brain are briefly discussed as an application.
- Chair: Yasuyuki Nogami (Okayama University)
- Speaker: Naofumi Homma (Tohoku University)
- Title: Recent topics on hardware security
- Abstract: Hardware security in mobile and embedded systems is drawing much attention in the context of the rapid growth of Internet-of-Things. Due to the easier accessibility, security threats and vulnerabilities for “things” located everywhere are more critical in comparison with PCs and servers in a room. In particular, the threats of side-channel attacks are non-trivial because they can be done by relatively low-cost equipment in a non-destructive manner. In the last few decades, a variety of side-channel attacks have been introduced and defeated, but they are still being one of the hottest topics in the field of hardware security research. This talk will start with an overview of side-channel attacks and their countermeasures, and then introduce the-state-of-the-art studies including a novel reactive countermeasure that makes it possible to prevent all the microprobe-based side-channel attacks.
- Chair: Yasuhiko Nakashima (NAIST)
- Speaker: Ryousei Takano (AIST)
- Title: Flow Centric Computing Platform
- Abstract: Some innovative data center architectures have been proposed from both academia and industry, for instance, Open Compute Project, UCB ASPIRE FireBox, Intel Rack scale architecture, and HP the machine. “Disaggregation” is a keyword of such architectures. Computers are divided into their component parts and a rack-scale or data center-scale computer is re-aggregated from the resource pool to meet users' requirements. Optical network is a promising technology to connect components with huge bandwidth and low power consumption. To utilize such advantages from applications, there are research opportunities for system software. This talk presents a new computing paradigm called “flow centric computing,” and its platform technologies including data center architecture and system software.
- Chair: Takashi Yokota (Utsunomiya University)
- Speaker: Kentaro Sano (Tohoku University)
- Title: FPGA-based Custom Computing : A Challenge to Architectural Exploration
- Abstract: The way of computing is being changed due to the end of multicore scaling with traditional microprocessor architectures. The issues of the dark silicon, the bandwidth and latency of data movement, a lack of parallelism, and a general-purpose structure make the multicore architectures inefficient in terms of area, energy, and cost, while they still keep supports from users for their software programming model. FPGA-based custom computing has been spotlighted as a promising means to make a breakthrough in computing efficiency. By customizing hardware for the physical constrains and target algorithms, we can provide a solution to achieve efficient computation with FPGAs. However, the customization is simultaneously a big challenge to computational architectures. We have to find the best mix of a static part (hardware) and a dynamic part (controls) with their interfaces specialized for each individual problem. We also need to improve productivity for architectural exploration and hardware design. This talk presents the advantages, issues, and research opportunities of custom computing with its recent topics.
- Chair: Susumu Matsumae (Saga University)
- Speaker: Toshimitsu Masuzawa (Osaka University)
- Title: Self-stabilizing distributed algorithms
- Abstract: Self-stabilization is a general paradigm to provide forward recovery capabilities to distributed systems. A self-stabilizing distributed algorithm can eventually recover its intended behavior even when starting from an arbitrary initial configuration, and thus, it has high adaptability to transient faults (e.g., process state corruptions and message corruptions) and network topology changes. In this talk, we introduce fundamentals of self-stabilizing distributed algorithms and present some of current trends of the research.
- Chair: Hiroyuki Takizawa (Tohoku University)
- SPEAKER: Mary Hall (Professor, University of Utah)
- TITLE: The Role of Compiler Optimization and Autotuning for Reducing Data Movement in High-Performance Applications
- ABSTRACT:As the cost of moving data on current and future architectures becomes increasingly dominant, the challenges of developing high-performance applications are increasingly onerous. The goal of compiler optimization in high-performance computing is to take as input a computation that is architecture independent and maintainable and produce as output efficient implementations of the computation tuned for the target architecture. Autotuning empirically evaluates a search space of possible implementations of a computation to identify the implementation that best meets its optimization criteria (e.g., performance, power, or both). Combining the two concepts, autotuning compilers generate this search space of highly-tuned implementations either automatically or with programmer guidance. This talk will explore the role of compiler technology in achieving very high levels of performance, comparable to what is obtained manually by experts. As a case study, it will highlight some of the aggressive optimizations required to reduce communication for a specific high-performance application domain that is notoriously memory bound: geometric multigrid and the stencil computations within them.
- Chair: Nobuo Funabiki (Okayama University)
- Speaker: Masakatsu Morii (Kobe University)
- TITLE: On vulnerability of SSL/TLS and that implementation; Attacks on Broadcast RC4 and others
- ABSTRACT: RC4 is adopted in many software applications and standard protocols such as SSL/TLS, WEP, Microsoft Lotus, Oracle secure SQL and more. It is discussed that the practical security of RC4 in broadcast setting where the same plaintext is encrypted with different user keys. For SSL/TLS using initial bytes of the keystream, the broadcast setting is converted into the multi-session setting where the target plaintext block are repeatedly sent in the same position in the plaintexts in multiple SSL/TLS sessions. Furthermore we discuss the some cache timing attacks of the implementation on OpenSSL. These attacks exploit the cache hits and misses that occur during the encryption process and recovers full AES key by measuring the total execution time of an encryption, named “CREAM”. Lastly, a state of CAESAR: Competition for Authenticated Encryption: Security, Applicability, and Robustness is presented with my personal impressions.