DOCSLIB.ORG

Physics Illinois News

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Physics Illinois News PHYSICS ILLINOIS NEWS THE DEPARTMENT OF PHYSICS AT THE UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN • 2006 NUMBER 1 Quantum Computing without “Computing” BY JAMES E. KLOEPPEL computer did not run (O. Hosten, et al., Nature 439, 949–952 [2006]). y combining quantum Quantum computers have the Bcomputation and quantum potential for solving certain types of interrogation, Bardeen Professor of problems much faster than classical Physics Paul Kwiat and his group have computers. Speed and efficiency are found an exotic way of determining gained because quantum bits can be an answer to an algorithm—without placed in superpositions of 1 and 0, ever running the algorithm. Using an as opposed to classical bits, which optical-based quantum computer, the are either 1 or 0. Moreover, the logic researchers have presented the first behind the coherent nature of quantum demonstration of “counterfactual information processing often deviates computation,” inferring information from intuitive reasoning, leading to about an answer, even though the some surprising effects. “It seems absolutely bizarre that counterfactual computation—using tauffer information that is counter to what The Kwiat research group at the Alma Mater rian S must have actually happened—could y L. B find an answer without running the not run the search algorithm,” said interrogation schemes, and to my mind, entire quantum computer,” said Kwiat. graduate student Onur Hosten, lead quantum mechanics doesn’t get any hoto b P “But the nature of quantum author of the Nature paper. “We also more mysterious than this.” interrogation makes this amazing showed theoretically how to obtain While the researchers’ optical feat possible.” the answer without ever running the quantum computer cannot be scaled Sometimes called interaction-free algorithm, by using a ‘chained Zeno’ up, using these kinds of interrogation measurement, quantum interrogation effect.” techniques may make it possible to is a technique that makes use of Through clever use of beam splitters reduce errors in quantum computing, wave-particle duality (in this case, of and both constructive and destructive Kwiat said. “Anything you can do to photons) to search a region of space interference, the researchers can put reduce the errors will make it more likely without actually entering that region each photon in a superposition of that eventually you’ll get a large-scale of space. Utilizing two coupled optical taking two paths. Although a photon quantum computer.” interferometers, nested within a can occupy multiple places simul- In addition to Kwiat and Hosten, third, Kwiat’s team succeeded in taneously, it can make an actual co-authors of the Nature paper are Paul Kwiat, right, John Bardeen Professor counterfactually searching a four- appearance at only one location. graduate students Julio Barreiro, of Physics and Electrical and Computer element database using Grover’s Its presence defines its path, and that Nicholas Peters, and Matthew Rakher Engineering, and graduate student quantum search algorithm. can, in a very strange way, negate the (BS ’03, now at the University of Onur Hosten have shown that quantum “By placing our photon in a need for the search algorithm to run. California, Santa Barbara). I computation can determine the answer quantum superposition of running “In a sense, it is the possibility that This work was funded by the Disruptive Technologies to an algorithm—without ever running and not running the search algorithm, the algorithm could run that prevents the algorithm. Office and the National Science Foundation. The we obtained information about the the algorithm from running,” Kwiat conclusions presented are those of the authors and answer even when the photon did said. “That is at the heart of quantum not necessarily of the funding agencies. Of Bits and Bytes: Physics Pioneers Computing at Illinois n 1949, the University received In 1964, the Digital Computer the University of Illinois and beyond. Iapproval to build a copy of a Laboratory—comprised mainly of This issue of Physics Illinois News computer being designed at the Physics faculty—was renamed the features interviews with several of the Institute for Advanced Study in Department of Computer Science, central figures in computing history Princeton, New Jersey. The Illinois but the undergraduate degree program at the U of I, and we invite you to computer group was organized as the in computer science wasn’t introduced enjoy this glimpse of our important Digital Computer Laboratory with until 1972. heritage. I Ralph Meagher, a physicist, named In 1985, the National Science chief engineer for the computer project. Foundation established the Center Completed in 1951 and named for Supercomputing Research and ORDVAC, the computer was one of Development and the National Center the fastest in existence—performing for Supercomputing Applications, with an add time of 92 ms and a Abraham Traub and Ralph Meagher with ILLIAC I making Urbana-Champaign the multiplication time of 700 ms. nation’s only university campus with Measuring 10-ft long, 2-ft wide the University’s radio telescope. two supercomputer centers. Larry and 8.5-ft high, the 5-ton machine Using the ILLIAC for brain power, Smarr, an astrophysicist, was named contained 2,800 vacuum tubes. U of I faculty introduced PLATO, the NCSA’s first director. Its successor, dubbed ILLIAC, was nation’s first computer-based program The NCSA would drive the completed in Fall 1952 and was one of instruction. Beginning in 1959, development of computer visualization of the first electronic computers built Physics faculty member and Control software. Physics graduate student and owned entirely by an educational Systems Laboratory Director Daniel Brand Fortner (BS ’77, MS ’82, PhD institution. Operating initially only Alpert championed the idea of Chalmers ’93), co-founder of Spyglass Inc., 8 hours a day, its schedule was Sherwin (also a Physics faculty member) licensed the software from the gradually expanded to 24-hours-per- that computers could be used for University and developed the first day, six-days-a-week operation. ILLIAC instruction, a concept that was to graphical browser for the Internet, would provide the major portion of the emerge as PLATO. About 25 years later, Mosaic®, in 1995. Physics faculty developed a pattern recognition computer services for the University for Lorella Jones would write the country’s The Department of Physics—its computer, which would become ILLIAC III, to 10 years, finding use in the design of first computerized homework problems faculty and alumni—has had an solve the problem of interpreting millions of other research instruments such as on PLATO for an elementary physics enormous influence on the bubble-chamber photographs. high-energy particle accelerators and course at Illinois. development of computing at 2 PHYSICS ILLINOIS NEWS • 2006 NUMBER 1 Message from the Head his is the seventh issue of Physics only 20 percent of the total university system levels will be a TIllinois News and my last as head UIUC budget comes from state multi-year challenge. I am confident of the department. I will be stepping appropriations, a smaller fraction than that the Department of Physics will down as head and retiring from the comes from tuition. Nevertheless, we come out stronger in the end, but it University this summer. At that time, have been able to hire new faculty, the will need help from all of its friends. I will have completed 39 years as a lifeblood of our department, during I am also confident that the member of the faculty and 6 as head. every year that I have been head. Department will play a major role in As I look back, I am amazed how Between January 2005, when Joe assisting the College of Engineering quickly 39 years has passed and how White took office as the new president and the University of Illinois at much the personnel, research, of the University of Illinois system, Urbana-Champaign to meet their teaching styles, and the outreach and Summer 2006, when my challenges and goals as well. I also programs of the Department of successor takes over, there will have know that the new head of the Physics have changed, and at the been a complete turnover of the Department of Physics will struggle same time how the ethos of administrative hierarchy: Head of with the same issues that Wheeler department—the collegial Urbana Physics, Dean of Engineering, Provost, Loomis and every department head spirit—has endured. The latter is the Chancellor, and President. Starting in since have had to deal with—money most important news I have to Spring 2005, the ground work was and space! report, and I am proud of that. laid for a strategic planning process at It has been a great honor to have During my tenure as head, all levels of the University to redefine served as head of the Department of especially in the last four years, the its mission and its relationship to the Physics, and I am deeply humbled to long-standing implicit covenant State and, at the same time, to expand have been entrusted with that between public universities and its national and international reach. opportunity and responsibility. I their state governments has changed The planning process will come to irreversibly, even at land grant completion in June 2006. universities such as ours. In the Executing the strategic plans at the current academic year, for example, departmental, college, campus, and Jeremiah D. Sullivan First Atom-by-Atom Simulation of a Life Form he computing horsepower of one collaborated with crystallographers at Tof the world’s most powerful the University of California, Irvine— supercomputers has been harnessed Alexander McPherson, a professor of by Swanlund Professor of Physics molecular biology and biochemistry, Klaus Schulten and his research group and research specialist Steven Larson. to visualize the behavior of a complete The group’s results were published in life form, the satellite tobacco mosaic the March issue of Structure (P.L.
Load more

Recommended publications
  • Famous Physicists Himansu Sekhar Fatesingh
    Fun Quiz FAMOUS PHYSICISTS HIMANSU SEKHAR FATESINGH 1. The first woman to 6. He first succeeded in receive the Nobel Prize in producing the nuclear physics was chain reaction. a. Maria G. Mayer a. Otto Hahn b. Irene Curie b. Fritz Strassmann c. Marie Curie c. Robert Oppenheimer d. Lise Meitner d. Enrico Fermi 2. Who first suggested electron 7. The credit for discovering shells around the nucleus? electron microscope is often a. Ernest Rutherford attributed to b. Neils Bohr a. H. Germer c. Erwin Schrödinger b. Ernst Ruska d. Wolfgang Pauli c. George P. Thomson d. Clinton J. Davisson 8. The wave theory of light was 3. He first measured negative first proposed by charge on an electron. a. Christiaan Huygens a. J. J. Thomson b. Isaac Newton b. Clinton Davisson c. Hermann Helmholtz c. Louis de Broglie d. Augustin Fresnel d. Robert A. Millikan 9. He was the first scientist 4. The existence of quarks was to find proof of Einstein’s first suggested by theory of relativity a. Max Planck a. Edwin Hubble b. Sheldon Glasgow b. George Gamow c. Murray Gell-Mann c. S. Chandrasekhar d. Albert Einstein d. Arthur Eddington 10. The credit for development of the cyclotron 5. The phenomenon of goes to: superconductivity was a. Carl Anderson b. Donald Glaser discovered by c. Ernest O. Lawrence d. Charles Wilson a. Heike Kamerlingh Onnes b. Alex Muller c. Brian D. Josephson 11. Who first proposed the use of absolute scale d. John Bardeen of Temperature? a. Anders Celsius b. Lord Kelvin c. Rudolf Clausius d.
    [Show full text]
  • Emerging Physics a Fresh Approach to Viewing the Complexity of the Universe
    GETTY IMAGES Emerging physics A fresh approach to viewing the complexity of the Universe. A Different Universe: Reinventing (an analogy that Robert Laughlin draws from with those of life,the biomolecules,for which Physics from the Bottom Down Christina Rossetti’s poem Who Has Seen the he has considerable admiration.Then we hear by Robert Laughlin Wind?). Our identity and perceptions are his own ideas on biology, which will not be Basic Books: 2005. 304 pp. $26, £19.99 all the collective behaviour of ‘ghosts’, who to everyone’s taste but are certainly thought- Philip Anderson borrow their reality from each other. provoking. Finally, his view of complexity Laughlin gives the reader a quick tour science surprised and pleased me with its I should make my interests clear right at the through much of physics (without a single relative benevolence. start. For many years I have thought that a equation). There is a slight emphasis on the Despite the above fulsome praise, this is book such as this should be written, and quantum theory of condensed matter, in so not by any means a perfect book, even for its have been urged to write it myself. I didn’t far as it explains such things as computers purpose.Laughlin is not reliably careful with do so, and couldn’t possibly have written (with a sceptical side glance at quantum facts, whether scientific or historical. For one as suited as this is for its target audience. computation), the properties of ordinary example, it has rhetorical value to give his A Different Universe is a book about what metals, and the like.
    [Show full text]
  • Contention-Free Periodic Message Scheduler Medium Access Control in Wireless Sensor / Actuator Networks
    Contention-Free Periodic Message Scheduler Medium Access Control in Wireless Sensor / Actuator Networks Thomas W. Carley Moussa A. Ba Rajeev Barua David B. Stewart ECE Department Embedded Research ECE Department Embedded Research University of Maryland Solutions University of Maryland Solutions [email protected] mba@embedded- [email protected] dstewart@embedded- zone.com zone.com Abstract actuator network applications have timeliness constrains, both on activities on individual nodes, and across multiple This paper presents a time division multiple access nodes. This results in timeliness constraints on both tasks medium access control protocol for wireless sensor / running on nodes and messages exchanged by nodes. actuator networks implemented with a contention-free Medium access control (MAC) is one of the most message scheduler. A message scheduler is used to important, and most studied, protocol layers in wireless determine which message has access to the medium at sensor / actuator networks. Wireless sensor / actuator any time. A set of messages is contention-free if only one networks are often characterized by extreme constraints message is ready at a time. Otherwise, some other on size, cost, power, and timeliness. This imposes criterion such as priority must be used to resolve constraints on every aspect of the design of these contention. In this case the message scheduler at each networks, especially the protocol stack. Wireless node in the network must schedule all messages in order networks are inherently broadcast media; all nodes in the to resolve contention. If the schedule is contention-free network share one common communication medium. however, then each node only schedules the messages that Therefore, a method for resolving contention when interest it.
    [Show full text]
  • 2017-IMSA-Fund-Annual-Report.Pdf
    1 from the fund board president Jacob Plummer ’96 IMSA Fund Board President A few years ago, a past president of the IMSA Fund Board, John Hoesley, called After joining the board, I learned for the first time of thousands of professional me and said “I’d like you to join us.” I told him “I’ve never heard of the Board – development workshops led by Dr. Storm Robinson’s outreach division at what do you do?” And he said, “We raise money, we open doors, and we support IMSA - impacting students and teachers across Illinois. IMSA.” Like all of us, I am grateful for the funding the State of Illinois provides It was easy to say yes. As an alum, many of my closest friends are people I met for IMSA. Carl Sagan said IMSA was a gift from the people of Illinois on campus. And, IMSA gave me opportunities I hadn’t even imagined. I joined to the human future – and it is. But our community has a role too. The the Board out of gratefulness. However, I’ve stayed on the board for two other contributions of all our donors, our Chicago companies and foundations reasons and these are reasons that might also matter to you. – great supporters like Ball Horticultural Company, Boeing, BP, Caterpillar Foundation, ComEd, Dart Foundation, EcoLab, Hansen-Furnas Foundation, The best thing about joining the board was having a way to connect with the Harris Family Foundation, Nicor Gas, NOAA, Pentair, Sodexo, and Tellabs Academy. Today, I regularly meet students and faculty who have incredible Foundation to name just a few.
    [Show full text]
  • 2019 Annual Report
    BECKMAN CENTER 279 Campus Drive West Stanford, CA 94305 650.723.8423 Stanford University | Beckman Center 2019 Annual Report Annual 2019 | Beckman Center University Stanford beckman.stanford.edu 2019 ANNUAL REPORT ARNOLD AND MABEL BECKMAN CENTER FOR MOLECULAR AND GENETIC MEDICINE 30 Years of Innovation, Discovery, and Leadership in the Life Sciences CREDITS: Cover Design: Neil Murphy, Ghostdog Design Graphic Design: Jack Lem, AlphaGraphics Mountain View Photography: Justin Lewis Beckman Center Director Photo: Christine Baker, Lotus Pod Designs MESSAGE FROM THE DIRECTOR Dear Friends and Trustees, It has been 30 years since the Beckman Center for Molecular and Genetic Medicine at Stanford University School of Medicine opened its doors in 1989. The number of translational scientific discoveries and technological innovations derived from the center’s research labs over the course of the past three decades has been remarkable. Equally remarkable have been the number of scientific awards and honors, including Nobel prizes, received by Beckman faculty and the number of young scientists mentored by Beckman faculty who have gone on to prominent positions in academia, bio-technology and related fields. This year we include several featured articles on these accomplishments. In the field of translational medicine, these discoveries range from the causes of skin, bladder and other cancers, to the identification of human stem cells, from the design of new antifungals and antibiotics to the molecular underpinnings of autism, and from opioids for pain
    [Show full text]
  • Open Dissertation Draft Revised Final.Pdf
    The Pennsylvania State University The Graduate School ICT AND STEM EDUCATION AT THE COLONIAL BORDER: A POSTCOLONIAL COMPUTING PERSPECTIVE OF INDIGENOUS CULTURAL INTEGRATION INTO ICT AND STEM OUTREACH IN BRITISH COLUMBIA A Dissertation in Information Sciences and Technology by Richard Canevez © 2020 Richard Canevez Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy December 2020 ii The dissertation of Richard Canevez was reviewed and approved by the following: Carleen Maitland Associate Professor of Information Sciences and Technology Dissertation Advisor Chair of Committee Daniel Susser Assistant Professor of Information Sciences and Technology and Philosophy Lynette (Kvasny) Yarger Associate Professor of Information Sciences and Technology Craig Campbell Assistant Teaching Professor of Education (Lifelong Learning and Adult Education) Mary Beth Rosson Professor of Information Sciences and Technology Director of Graduate Programs iii ABSTRACT Information and communication technologies (ICTs) have achieved a global reach, particularly in social groups within the ‘Global North,’ such as those within the province of British Columbia (BC), Canada. It has produced the need for a computing workforce, and increasingly, diversity is becoming an integral aspect of that workforce. Today, educational outreach programs with ICT components that are extending education to Indigenous communities in BC are charting a new direction in crossing the cultural barrier in education by tailoring their curricula to distinct Indigenous cultures, commonly within broader science, technology, engineering, and mathematics (STEM) initiatives. These efforts require examination, as they integrate Indigenous cultural material and guidance into what has been a largely Euro-Western-centric domain of education. Postcolonial computing theory provides a lens through which this integration can be investigated, connecting technological development and education disciplines within the parallel goals of cross-cultural, cross-colonial humanitarian development.
    [Show full text]
  • Slate XNS TSN Network Scheduler with Browser-Based Topology Modeling
    Slate XNS TSN Network Scheduler with Browser-based Topology Modeling Key Benefits Intuitive graphical topology modeling Incremental scheduling of new components and data streams Open, standard configuration for any TSN device One-click scheduling for complex networks Optimized bandwidth use for non-scheduled traffic Slate XNS is a browser-based, user-friendly software that makes it easy to model topologies, create schedules and deploy configurations for TSN networks. Offline network configuration is made possible by the intuitive GUI which provides a topology view or table-based editor for managing components and data streams. Schedules are calculated with just one click via TTTech’s built-in scheduling engine, and network components are configured using open, standard YANG models. Topology Modeling Network Scheduler Graphical One-Click Scheduling The graphical view is ideal for modeling small and TTTech’s high performance scheduling engine is medium sized network topologies. New components built-in to the Slate XNS software. After building the can be added by dragging and dropping into the topology and defining the data streams, the schedule topology. Data streams can also be represented as can be created with just one click. Schedules can logical connections between components. even be updated incrementally when new components or data streams are added. Table-Based Slate XNS schedules IEEE 802.1Qbv and 802.1Qbu The table editor is designed for modeling large traffic as well as supporting cut-through data network topologies and mass editing. Components, streams. Scheduling can be adjusted to optimize data streams and other parameters can be inputted bandwidth for non-scheduled traffic.
    [Show full text]
  • William Bradford Shockley
    William Bradford Shockley Born February 13, 1910, London, UK- died 1989, Santa Clara, Calif.; with Walter Brattain and John Bardeen, inventor of the transistor in 1947, the 1956 Nobel laureate. Education: BS, physics, California Institute of Technology, 1932; PhD, physics, MIT, 1936. Professional Experience: Bell Telephone Laboratories: member, Technical Staff, 1936- 1942 and 1945-1954, director, Transistor Physics Research Facility, 1954; director of research, Antisubmarine Warfare Operations Research Group, US Navy, 1942-1944; founder, Shockley Semiconductor Laboratory, 1954-1989; Stanford University: lecturer, 1958-1963, Alexander M. Poniatoff Professor of Engineering Science and Applied Science, 1963-1975, professor emeritus, 1975-1989. Honors and Awards: Nobel Prize in physics,1 1956. Co-inventor of the transistor in 1947 with John Bardeen and Walter Brattian, Shockley participated in one of the most important discoveries of the century. They applied for a patent in 1948; this device was described as a germanium “transfer resistance” unit, from which the name “transistor” was derived. Shockley continued his research on the device to create the germanium junction transfer transistor, which was much more reliable than the first unit. From this start he founded Shockley Semiconductor Laboratories in Santa Clara Valley in 1954. After he received the Nobel Prize in 1956, disenchantment with Shockley's management style and his propensity for pure research led to the defection of the “Fairchild Eight” in 1957, and the deterioration of his company. His controversial views on genetics and his racist theories have shocked the society around him, but he has continued his research into “grave world problems.” QUOTATION “The only heritage I can leave to Billy is the feeling of power and joy of responsibility for setting the world right on something.” (Shockley's mother, about her 8-year-old son) BIBLIOGRAPHY Biographical 1 Jointly with John Bardeen and Walter H.
    [Show full text]
  • National Centre for Biological Sciences
    Cover Outside Final_452 by 297.pdf 1 16/01/19 10:51 PM National Centre for Biological Sciences Biological for National Centre National Centre for Biological Sciences Tata Institute of Fundamnetal Research Bellary Road, Bangalore 560 065. India. P +91 80 23666 6001/ 02/ 18/ 19 F +91 80 23666 6662 www.ncbs.res.in σ 2 ∂ NCBSLogo — fd 100 ∂t rt 100 fd 200 10-9m 10-5m 10-2m 1m ANNUAL REPORT 2017-18 REPORT ANNUAL National Centre for Biological Sciences ANNUAL REPORT 2017-18 Cover Inside Final_452 by 297.pdf 1 12/01/19 7:00 PM National Centre for Biological Sciences ANNUAL REPORT 2017-18 A cluster of gossamer-winged dragonflies from Dhara Mehrotra’s exhibition “Through Clusters and Networks”, as a part of the TIFR Artist-in-Residence programme PHOTO: DHARA MEHROTRA CONTENTS Director’s Map of Research 4 7 Note Interests Research Theory, Simulation, New Faculty 8 and Modelling of 88 Reports SABARINATHAN RADHAKRISHNAN 8 Biological Systems SHACHI GOSAVI · MUKUND THATTAI · SANDEEP KRISHNA · MADAN RAO · SHASHI THUTUPALLI Biochemistry, Biophysics, 20 and Bioinformatics JAYANT UDGAONKAR · M K MATHEW R SOWDHAMINI · ASWIN SESHASAYEE RANABIR DAS · ARATI RAMESH · ANJANA BADRINARAYANAN · VINOTHKUMAR K R Cellular Organisation 38 and Signalling SUDHIR KRISHNA · SATYAJIT MAYOR · RAGHU PADINJAT · VARADHARAJAN SUNDARAMURTHY 48 Neurobiology UPINDER S BHALLA · SANJAY P SANE · SUMANTRA CHATTARJI · VATSALA THIRUMALAI · HIYAA GHOSH 60 Genetics and Development K VIJAYRAGHAVAN · GAITI HASAN P V SHIVAPRASAD · RAJ LADHER · DIMPLE NOTANI 72 Ecology and Evolution MAHESH
    [Show full text]
  • Hybrid Event-Time-Triggered Networked Control Systems: Scheduling-Event
    Information Sciences 305 (2015) 269–284 Contents lists available at ScienceDirect Information Sciences journal homepage: www.elsevier.com/locate/ins Hybrid event-time-triggered networked control systems: Scheduling-event-control co-design ⇑ Shixi Wen a, Ge Guo a, , Wing Shing Wong b a School of Information Science and Technology, Dalian Maritime University, Dalian 116026, China b Department of Information Engineering, The Chinese University of Hong Kong, Hong Kong, China article info abstract Article history: This paper is concerned with the simultaneous stabilization of a collection of plants Received 19 August 2014 remotely controlled via a wireless network. The so-called most regular binary sequences Received in revised form 21 December 2014 (MRBSs) are used to allocate the limited network channels. To further reduce unnecessary Accepted 27 January 2015 transmissions, an event-triggered transmitter is introduced at each plant. Under the hybrid Available online 7 February 2015 event-time-triggered transmission strategy, a sufficient stability condition is derived for each plant. Then, a unified co-design framework is presented with which the scheduling Keywords: functions (i.e., the MRBSs), the event-triggering condition and the controller gains can be Networked control systems determined simultaneously. The main results are established using piecewise Lyapunov Capacity limitation Event-triggered transmission functional method and average dwell time technique. The effectiveness of the co-design Protocol sequence method is demonstrated by a numerical example. Scheduling-event-control co-design Ó 2015 Elsevier Inc. All rights reserved. 1. Introduction In wireless networked systems (such as wireless sensor networks) consisting of a large number of nodes, network capa- city limitation is often a serious issue.
    [Show full text]
  • Transforming Microtubules Into Microshuttles for Advanced Immunoassay Applications
    Transforming Microtubules into Microshuttles for Advanced Immunoassay Applications by Jenna Marie Campbell A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Mechanical Engineering) in the University of Michigan 2014 Doctoral Committee: Professor Edgar Meyhöfer, Chair Assistant Professor Barry Grant Assistant Professor Ajit P. Joglekar Professor Katsuo Kurabayashi © Jenna Campbell 2014 All Rights Reserved For Lisa Orr and Vivek Tomer— For pushing me forward every time I wanted to take a step back. Your love, support, and guidance have been invaluable. ii ACKNOWLEDGEMENTS This dissertation is the culmination of many years of work, which would not have been possible without the guidance of my advisor, Edgar Meyhöfer. He challenged me, taught me how to ask questions, and how to “do science.” Always having an open door, your support and expertise has been invaluable academically, professionally, and personally. I would also like to acknowledge Katsuo Kurabayashi for your expertise and support. Our meetings gave me a broader perspective on our projects, which played a large role in shaping my academic and professional goals. Committee members Ajit Joglekar and Barry Grant have always been generous with their time, and I am thankful for the advice and expertise they have given throughout my graduate career. My co-workers Charles Jiang and Neha Kaul played a critical role in this work. Thank you for bringing laughter along with great insight into my work. I would like to acknowledge Professor James Gole for encouraging me to continue my education, which brought me to the University of Michigan. I would not have discovered my love of science and research if it were not for your passion to inspire undergraduates and educate young scientists.
    [Show full text]
  • Graphpulse: an Event-Driven Hardware Accelerator for Asynchronous Graph Processing
    2020 53rd Annual IEEE/ACM International Symposium on Microarchitecture (MICRO) GraphPulse: An Event-Driven Hardware Accelerator for Asynchronous Graph Processing Shafiur Rahman Nael Abu-Ghazaleh Rajiv Gupta Computer Science and Engineering Computer Science and Engineering Computer Science and Engineering University of California, Riverside University of California, Riverside University of California, Riverside Riverside, USA Riverside, USA Riverside, USA [email protected] [email protected] [email protected] Abstract—Graph processing workloads are memory intensive GraphLab [31], GraphX [16], PowerGraph [17], Galois [37], with irregular access patterns and large memory footprint result- Giraph [4], GraphChi [28], and Ligra [49]. ing in low data locality. Their popular software implementations Large scale graph processing introduces a number of chal- typically employ either Push or Pull style propagation of changes through the graph over multiple iterations that follow the Bulk lenges that limit performance on traditional computing systems. Synchronous Model. The performance of these algorithms on First, memory-intensive processing stresses the memory system. traditional computing systems is limited by random reads/writes Not only is the frequency of memory operations relative of vertex values, synchronization overheads, and additional over- to compute operations high, the memory footprints of the heads for tracking active sets of vertices or edges across iterations. graph computations are large. Standard techniques in modern In this paper, we present GraphPulse, a hardware framework for asynchronous graph processing with event-driven schedul- processors for tolerating high memory latency, such as caching ing that overcomes the performance limitations of software and prefetching, have limited impact because irregular graph frameworks. Event-driven computation model enables a parallel computations lack data locality.
    [Show full text]