792 research outputs found
Overview of the SpiNNaker system architecture
Get PDFSpiNNaker (a contraction of Spiking Neural Network Architecture) is a million-core computing engine whose flagship goal is to be able to simulate the behaviour of aggregates of up to a billion neurons in real time. It consists of an array of ARM9 cores, communicating via packets carried by a custom interconnect fabric. The packets are small (40 or 72 bits), and their transmission is brokered entirely by hardware, giving the overall engine an extremely high bisection bandwidth of over 5 billion packets/s. Three of the principle axioms of parallel machine design - memory coherence, synchronicity and determinism - have been discarded in the design without, surprisingly, compromising the ability to perform meaningful computations. A further attribute of the system is the acknowledgement, from the initial design stages, that the sheer size of the implementation will make component failures an inevitable aspect of day-to-day operation, and fault detection and recovery mechanisms have been built into the system at many levels of abstraction. This paper describes the architecture of the machine and outlines the underlying design philosophy; software and applications are to be described in detail elsewhere, and only introduced in passing here as necessary to illuminate the description
SpiNNaker: a 1-W 18-core system-on-chip for massively parallel neural network simulation
Get PDFThe modelling of large systems of spiking neurons is computationally very demanding in terms of processing power and communication. SpiNNaker is a massively-parallel computer system designed to model up to a billion spiking neurons in real time. The basic block of the machine is the SpiNNaker multicore System-on-Chip, a Globally Asynchronous Locally Synchronous (GALS) system with 18 ARM968 processor nodes residing in synchronous islands, surrounded by a light-weight, packet-switched asynchronous communications infrastructure. The MPSoC contains 100 million transistors in a 102 mm2 die, provides a peak performance of 3.96 GIPS and has a power consumption of 1W at 1.2V when all processor cores operate at nominal frequency. SpiNNaker chips were delivered in May 2011, were fully operational, and met power and performance requirement
Microbial enrichment culture responsible for the complete oxidative biodegradation of 3‑Amino-1,2,4-triazol-5-one (ATO), the reduced daughter product of the insensitive munitions compound 3‑Nitro-1,2,4-triazol-5-one (NTO)
No full text3-Nitro-1,2,4-triazol-5-one (NTO) is one of the main ingredients of many insensitive munitions, which are being used as replacements for conventional explosives. As its use becomes widespread, more research is needed to assess its environmental fate. Previous studies have shown that NTO is biologically reduced to 3-amino-1,2,4-triazol-5-one (ATO). However, the final degradation products of ATO are still unknown. We have studied the aerobic degradation of ATO by enrichment cultures derived from the soil. After multiple transfers, ATO degradation was monitored in closed bottles through measurements of inorganic carbon and nitrogen species. The results indicate that the members of the enrichment culture utilize ATO as the sole source of carbon and nitrogen. As ATO was mineralized to CO₂, N₂, and NH₄⁺, microbial growth was observed in the culture. Co-substrates addition did not increase the ATO degradation rate. Quantitative polymerase chain reaction analysis revealed that the organisms that enriched using ATO as carbon and nitrogen source were Terrimonas spp., Ramlibacter-related spp., Mesorhizobium spp., Hydrogenophaga spp., Ralstonia spp., Pseudomonas spp., Ectothiorhodospiraceae, and Sphingopyxis. This is the first study to report the complete mineralization of ATO by soil microorganisms, expanding our understanding of natural attenuation and bioremediation of the explosive NTO.Journal ArticleFinal article publishe
Letter from June Kizu, National Coalition for Redress/Reparations, to Edmund D. Edelman, Supervisor, County of Los Angeles, January 26, 1982
No full textLetter from June Kizu, National Coalition for Redress/Reparations (NCRR), to Edmund D. Edelman, Supervisor, County of Los Angeles, inviting Edelman to join the Day of Remembrance event in Los Angeles, California.The Jim Matsuoka Nikkei for Civil Rights and Redress Collection includes brochures, meeting notes and agendas, publications, booklets, and other material related to the Nikkei for Civil Rights and Redress (NCRR), formally known as the National Coalition for Redress/Reparations. The National Coalition for Redress/Reparations was officially formed on July 12, 1980, and included members of the Los Angeles Community Coalition for Redress/Reparations (LACCRR), Japanese Community Progressive Alliance (JCPA), Tule Lake Committee, Nihonmachi Outreach Committee, the Asian/Pacific Student Union, and other members of the community. The material was collected by Jim Matsuoka, a founding member of the organization. Matsuoka also served on the board and was the treasurer. In addition to the NCRR material, the collection also contains event flyers and Day of Remembrance material. For issues of the Nikkei for Civil Rights and Redress newsletter "Banner" published after 2007, visit the NCRR website at https://ncrr-la.org/
A CMOS VLSI Implementation of an Asynchronous ALU
No full textA CMOS self-timed ALU has been developed as part of an asynchronous implementation of the ARM microprocessor. This unit exploits the data dependency inherent in many arithmetic operations to enable a small, simple ALU to deliver a mean performance comparable with that of a more sophisticated synchronous one with consequent reductions in both silicon area and electrical power consumption. The self-timed nature of the unit means that the majority of operations complete quickly whilst allowing rare `worst-case' operations to take longer, maintaining a high average throughput. This paper presents instruction usage statistics to justify the claimed performance and SPICE simulation results of measurements taken from the layout. Keyword Codes: B.2.0; B.2.1; B.7.1 Keywords: Arithmetic and Logic Structures, General; Arithmetic and Logic Structures, Design Styles,Integrated Circuits, Types and Design Styles 1. INTRODUCTION The ever-increasing availability of high performance processing has led ..
Adaptive pipeline structures for speculation control
No full textPipelining is a common method for improving the throughput of a system, especially when the majority of the processing is sequential. Unfortunately when the se-quentiality is broken, a pipelined system suffers additional delay and, most importantly for this work, energy waste which is roughly proportional to the pipeline depth. Standard pipelines cannot be modified once they are built so their depth is fixed. This paper proposes a method that allows the dynamic adaptation of the structure of an asynchronous pipeline, so that pipeline stages can be merged and split at run-time, allowing greater flexibility. It is based on novel latch controllers that can be configured dynamically as 'normal' or 'collapsed', i.e. keeping their latches permanently transparent. Using these controllers a model of AMULET3 was designed that is capable of changing its pipeline depth dynamically when branches are anticipated, in order to alleviate the energy loss when the branch finally arrives. © 2003 IEEE
A discontinuous Galerkin FEM multi-physics solver for the molten salt fast reactor
No full textNumerical simulations of fast MSRs constitute a challenging task. In fact, classical codes employed in reactor physics cannot be used, and new dedicated multi-physics tools must be developed, to capture the unique features of these systems: the strong coupling between neutronics and thermal-hydraulics due to the use of a liquid fuel, the effects on reactor kinetics induced by the precursors drift, the internal heat generation, and the shape of the core having no fuel pins as a repeated structure. In this work, we present a novel multi-physics tool being developed at TU Delft. The coupling is realized between an SN radiation transport code (PHANTOM-SN) and a RANS solver (DGFlows). Both in-house tools are based on a Discontinuous Galerkin Finite Element space discretization, characterized by local conservation, high-order accuracy, and allowing for high geometric flexibility. Implicit discretization in time is performed adopting Backward Differentiation Formulae. Cross sections are computed on an element base, starting from the local average temperature and a set of libraries generated at reference temperatures with Monte Carlo or deterministic codes. Comparison of the results obtained performing a suitable numerical benchmark created at LPSC/CNRS/Grenoble with those available in literature shows that the multi-physics tool is able to capture the unique phenomena characterizing fast liquid-fueled systems.Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.RST/Reactor Physics and Nuclear MaterialsRST/Radiation, Science and Technolog
An Adaptive Serial-Parallel CAM Architecture for Low-Power Cache Blocks
No full textThere is an on- oing debate about which consumes less energy: a RAMtagged associative cache with an intelligent order of accessing its tags and ways (e.g. way prediction), or a CAMtagged high associativity cache. If a CAM search can consume less than twice the energy of reading a tag RAM, it would probably be the preferred option for lowpower applications
- …
