词条 | SpiNNaker |
释义 |
| name = SpiNNaker: Spiking Neural Network Architecture | logo = | website = {{URL|http://apt.cs.manchester.ac.uk/projects/SpiNNaker/}} | commercial = | type = neuromorphic | location = Manchester, UK | owner = | founderimage = | founder = Steve Furber | date_of_establishment = | date_of_disestablishment = }}SpiNNaker (Spiking Neural Network Architecture) is a massively parallel, manycore supercomputer architecture designed by the Advanced Processor Technologies Research Group (APT) at the School of Computer Science, University of Manchester.[1] It is composed of 57,600 ARM9 processors (specifically ARM968), each with 18 cores and 128 MB of mobile DDR SDRAM, totalling 1,036,800 cores and over 7 TB of RAM.[2] The computing platform is based on spiking neural networks, useful in simulating the human brain (see Human Brain Project).[3][4][5][6][7][8][9][10][11] The completed design is housed in 10 19-inch racks, with each rack holding over 100,000 cores.[12] The cards holding the chips are held in 5 Blade enclosures, and each core emulates 1000 Neurons.[12] In total, the goal is to simulate the behavior of aggregates of up to a billion neurons in real time.[13] This machine requires about 100 kW from a 240 V supply and an air-conditioned environment.[14] SpiNNaker is being used as one component of the neuromorphic computing platform for the Human Brain Project.[15][16] On October 14, 2018 the HBP announced that the million core milestone had been achieved.[17][18] See also
References1. ^Advanced Processor Technologies Research Group {{Navboxes2. ^{{Cite web|url=http://apt.cs.manchester.ac.uk/projects/SpiNNaker/SpiNNchip/|title=SpiNNaker Project - The SpiNNaker Chip|last=|first=|date=|website=apt.cs.manchester.ac.uk|language=en|access-date=2018-11-17}} 3. ^{{citation |title=SpiNNaker Home Page |url=http://apt.cs.manchester.ac.uk/projects/SpiNNaker/ |publisher=University of Manchester |accessdate=11 June 2012}} 4. ^{{Cite journal | doi = 10.1109/JPROC.2014.2304638| title = The SpiNNaker Project| journal = Proceedings of the IEEE| volume = 102| issue = 5| pages = 652–665| year = 2014| last1 = Furber | first1 = S. B. | authorlink1 = Steve Furber| last2 = Galluppi | first2 = F. | last3 = Temple | first3 = S. | last4 = Plana | first4 = L. A. }} 5. ^{{Cite book| last1 = Xin Jin| last2 = Furber | first2 = S. B.| authorlink2 = Steve Furber| last3 = Woods | first3 = J. V.| doi = 10.1109/IJCNN.2008.4634194| chapter = Efficient modelling of spiking neural networks on a scalable chip multiprocessor |title = 2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence)| pages = 2812–2819| year = 2008| isbn = 978-1-4244-1820-6| pmid = | pmc = }} 6. ^A million ARM cores to host brain simulator News article on the project in the EE Times 7. ^{{Cite journal | last1 = Temple | first1 = S. | last2 = Furber | first2 = S. | authorlink2 = Steve Furber| doi = 10.1098/rsif.2006.0177 | title = Neural systems engineering | journal = Journal of the Royal Society Interface | volume = 4 | issue = 13 | pages = 193–206 | year = 2007 | pmid = 17251143| pmc = 2359843}} A manifesto for the SpiNNaker project, surveying and reviewing the general level of understanding of brain function and approaches to building computer modelof the brain. 8. ^{{Cite journal | last1 = Plana | first1 = L. A. | last2 = Furber | first2 = S. B. | authorlink2 = Steve Furber| last3 = Temple | first3 = S. | last4 = Khan | first4 = M. | last5 = Shi | first5 = Y. | last6 = Wu | first6 = J. | last7 = Yang | first7 = S. | doi = 10.1109/MDT.2007.149 | title = A GALS Infrastructure for a Massively Parallel Multiprocessor | journal = IEEE Design & Test of Computers | volume = 24 | issue = 5 | pages = 454 | year = 2007 | pmid = | pmc = }} A description of the Globally Asynchronous, Locally Synchronous (GALS) nature of SpiNNaker, with an overview of the asynchronous communications hardware designed to transmit neural 'spikes' between processors. 9. ^{{Cite book | doi = 10.1145/1542275.1542317| chapter = Understanding the interconnection network of SpiNNaker| title = Proceedings of the 23rd international conference on Conference on Supercomputing - ICS '09| pages = 286| year = 2009| last1 = Navaridas | first1 = J. | last2 = Luján | first2 = M. | last3 = Miguel-Alonso | first3 = J. | last4 = Plana | first4 = L. A. | last5 = Furber | first5 = S. | isbn = 9781605584980| citeseerx = 10.1.1.634.9481}} Modelling and analysis of the SpiNNaker interconnect in a million-core machine, showing the suitability of the packet-switched network for large-scale spiking neural network simulation. 10. ^{{Cite journal | last1 = Rast | first1 = A. | last2 = Galluppi | first2 = F. | last3 = Davies | first3 = S. | last4 = Plana | first4 = L. | last5 = Patterson | first5 = C. | last6 = Sharp | first6 = T. | last7 = Lester | first7 = D. | last8 = Furber | first8 = S. | authorlink8 = Steve Furber| doi = 10.1016/j.neunet.2011.06.014 | title = Concurrent heterogeneous neural model simulation on real-time neuromimetic hardware | journal = Neural Networks | volume = 24 | issue = 9 | pages = 961–978 | year = 2011 | pmid = 21778034 | pmc = }} A demonstration of SpiNNaker's ability to simulate different neural models (simultaneously, if necessary) in contrast to other neuromorphic hardware. 11. ^{{Cite journal | last1 = Sharp | first1 = T. | last2 = Galluppi | first2 = F. | last3 = Rast | first3 = A. | last4 = Furber | first4 = S. | authorlink4 = Steve Furber| title = Power-efficient simulation of detailed cortical microcircuits on SpiNNaker | doi = 10.1016/j.jneumeth.2012.03.001 | journal = Journal of Neuroscience Methods | volume = 210 | issue = 1 | pages = 110–118 | year = 2012 | pmid = 22465805 | pmc = }} Four-chip, real-time simulation of a four-million-synapse cortical circuit, showing the extreme energy efficiency of the SpiNNaker architecture 12. ^1 [https://www.youtube.com/watch?v=2e06C-yUwlc Video interview by computerphile with Steve Furber] 13. ^{{Cite web|url=http://apt.cs.manchester.ac.uk/projects/SpiNNaker/architecture/|title=SpiNNaker Project - Architectural Overview|last=|first=|date=|website=apt.cs.manchester.ac.uk|language=en|access-date=2018-11-17}} 14. ^{{Cite web|url=http://apt.cs.manchester.ac.uk/projects/SpiNNaker/hardware/|title=SpiNNaker Project - Boards and Machines|last=|first=|date=|website=apt.cs.manchester.ac.uk|language=en|access-date=2018-11-17}} 15. ^{{Cite journal | pmid = 24139655| year = 2013| author1 = Calimera| first1 = A| title = The Human Brain Project and neuromorphic computing| journal = Functional Neurology| volume = 28| issue = 3| pages = 191–6| last2 = Macii| first2 = E| last3 = Poncino| first3 = M| pmc = 3812737}} 16. ^{{Cite journal | doi = 10.1145/2601069| title = Neuromorphic computing gets ready for the (really) big time| journal = Communications of the ACM| volume = 57| issue = 6| year = 2014| last1 = Monroe | first1 = D. | pages = 13–15}} 17. ^{{Cite news|url=https://www.datacenterdynamics.com/news/spinnaker-brain-simulation-project-hits-one-million-cores-single-machine/|title=SpiNNaker brain simulation project hits one million cores on a single machine|access-date=2018-10-19|language=en}} 18. ^{{Citation|last=Petrut Bogdan|title=SpiNNaker: 1 million core neuromorphic platform|date=2018-10-14|url=https://www.youtube.com/watch?v=V3MlOAru6Qk|access-date=2018-10-19}} |list={{John McCarthy navbox}}{{philosophy of mind}}{{philosophy of science}}{{Evolutionary computation}}{{Computable knowledge}}{{Computer science}}{{Emerging technologies}}{{Robotics}} }}{{Philosophy of mind}}{{Cybernetics}}{{Computer science}}{{Evolutionary computation}}{{Emerging technologies}}{{Authority control}}{{comp-hardware-stub}} 12 : Artificial intelligence|Cybernetics|Formal sciences|Technology in society|Computational neuroscience|Emerging technologies|Unsolved problems in computer science|Computational fields of study|AI accelerators|Computer architecture|School of Computer Science, University of Manchester|Science and technology in Greater Manchester |
随便看 |
|
开放百科全书收录14589846条英语、德语、日语等多语种百科知识,基本涵盖了大多数领域的百科知识,是一部内容自由、开放的电子版国际百科全书。