# Register now to access 7 million high quality study materials (What's Course Hero?) Course Hero is the premier provider of high quality online educational resources. With millions of study documents, online tutors, digital flashcards and free courseware, Course Hero is helping students learn more efficiently and effectively. Whether you're interested in exploring new subjects or mastering key topics for your next exam, Course Hero has the tools you need to achieve your goals.

2 Pages

### S09-final-project-EDA-Luo

Course: ECE 498, Fall 2009
School: Lake County
Rating:

Word Count: 523

#### Document Preview

EDA Accelerating processes on the GPU Lijuan Luo UIUC Introduction In this project, I will try to accelerate two EDA process on the GPU, maze routing and lithography simulation. Maze routing is the fundamental routing algorithm in VLSI design. Current VLSI design needs to route tens of thousands of nets sequentially and the whole routing process may iterate multiple times owing to congestion. Hence, to accelerate...

Register Now

#### Unformatted Document Excerpt

Coursehero >> Illinois >> Lake County >> ECE 498

Course Hero has millions of student submitted documents similar to the one
below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.

Course Hero has millions of student submitted documents similar to the one below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.
EDA Accelerating processes on the GPU Lijuan Luo UIUC Introduction In this project, I will try to accelerate two EDA process on the GPU, maze routing and lithography simulation. Maze routing is the fundamental routing algorithm in VLSI design. Current VLSI design needs to route tens of thousands of nets sequentially and the whole routing process may iterate multiple times owing to congestion. Hence, to accelerate one net maze routing is very important. Lithography simulation is another well known time consuming process. I am targeting at 5 X speed up for maze routing and 20 X speedup for lithography simulation. Description Suppose there are two points A and B on a routing grid. The routing problem is to find the shortest path from A to B. In maze routing, we start from A and propagate to the neighboring grid cells step by step until we finally reach the target point B. In the CPU algorithm, a wave front is maintained for each step. At the beginning, only A is in the wave front. Then in each step all the grid cells neighboring the wave front cells are visited, and the newly visited cells will compose the new wave front for the next step. The parallelism of this algorithm is that each cell in the wave front can be handled in parallel. However, the difficulty is that we have to implement a queue structure on the GPU to record the wave front. The lithographic simulation method I am speeding up works in the following way: It decomposes the whole optical system into K coherent optical systems. Then the image density at a particular point (x,y) is calculated this by equation: I ( x, y ) = k | (O k )( x, y ) | 2 Here I ( x, y ) is the image density. k is a given constant. O( x, y ) is the object (field) function and k ( x, y ) is a function related with optical systems, which are both given. The symbol denotes 2D image convolution. Thus the way to speed up is very straightforward. The first kernel call is to calculate convolution in parallel. And then the second kernel call is to calculate the square of each element in parallel. k =1 K Objective Maze routing algorithm definitely has the bandwidth problem, and therefore we can only expect 5- 10 times speedup. Experiment shows that a simple implementation of the global queue structure can only bring 1.5X speedup. Thus I will explore some local queue structures, which will be merged and compacted to construct the global queue. The effect of this strategy is hard to estimate so far. Lithographic simulation is an easy speedup case and 20X speedup should be achieved. Background 1. knowledge about FFT, convolution 2. usage of scan primitive and texture memory in GPU programming 3. knowledge about EDA algorithms. Resources 1. Paper about lithography simulation: Jason Con and Yi Zou, Lithographic aerial image simulation with FPGA-based hardware acceleration, Proc. of 16th International Symposium on FPGA, pp. 67-76. 2. Papers for GPU-based graph algorithms: http://www.iiit.net/techreports/2007_103.pdf http://www.seas.upenn.edu/~kiderj/research/papers/APSP-gh08-fin-T.pdf Contact Information Lijuan Luo, lluo3@uiuc.edu Kishore Singal, chief scientist of Synopsys
Find millions of documents on Course Hero - Study Guides, Lecture Notes, Reference Materials, Practice Exams and more. Course Hero has millions of course specific materials providing students with the best way to expand their education.

Below is a small sample set of documents:

Lake County - ECE - 498
Adaptive Polynomial Transversal FilteringBei Peng and Seungchul Lee ECE, UIUCIntroductionThe scaling down of CMOS technology is brought up against analog circuit design, and settling these problems is a challenging prospective research area. The adapti
Lake County - ECE - 498
Parallel Programming in Finding Galaxy GroupsScott Kruger, Yiran Wang Astronomy Department IntroductionA very important part of cosmology is the analysis of the statistics of a density field traced by observable structures, such as distribution of galax
Lake County - ECE - 498
Substrate Coupling Analysis in Integrated CircuitsDae Hyun Kwon and Richard Tseng ECE, UIUC IntroductionAs clock and signal frequencies migrate to tens of GHz, coupled with a large number of circuit blocks encompassing digital, analog, and high-power ra
National Taiwan University - FIN - 726
Finance 726 MidtermProfessor Helwege Spring 2003You have the entire class period to finish this quiz. You may use a calculator, scrap paper, and a writing tool to complete the test. You may hand in the scrap paper (with your name on it) if you think it
National Taiwan University - FIN - 726
Finance 726 MidtermProfessor Helwege Spring 2003You have the entire class period to finish this quiz. You may use a calculator, scrap paper, and a writing tool to complete the test. You may hand in the scrap paper (with your name on it) if you think it
National Taiwan University - FIN - 726
Finance 726 MidtermProfessor Helwege Spring 2003You have the entire class period to finish this quiz. You may use a calculator, scrap paper, and a writing tool to complete the test. You may hand in the scrap paper (with your name on it) if you think it
National Taiwan University - FIN - 726
Finance 726 Midterm 1Professor Helwege Instructor Damsel Fall, 2003You have the entire class period to finish this exam. You may use a calculator, scrap paper, and a writing tool to complete this exam. You may hand in the scrap paper (with your name on
Gulf Coast Community College - ECE - 341
APPLICATION NOTEAPPLICATIONS OF MONOLITHIC BRIDGE DRIVERSHigh power monolithic bridge drivers are an attractive replacement for discrete transistors and half bridges in applications such as DC motor and stepper motor driving. This application guide desc
Colorado - ECE - 341
Application ReportSCPA032 - August 2001Improving System Interrupt Management Using the PCF8574 and PCF8574A I/O Expanders for the I 2C BusEmmanuel Tomdio Nana ABSTRACT In today's complex, embedded, computing and data-communication systems, interrupts a
Princeton - PHYS - 210
Yu Gan, Godfrey Miller, Carl Boettiger1. Reach Superfluid State 2. Basic observations 3. Observe Second Sound 4. Observe Angular Momentum Effects 5. Observe Fountain Effect1. Design plate 2. Run wires through brass connectors and fill connectors with ep
Princeton - PHYS - 210
Final Report on theMeasurement of the temperature of the Cosmic Microwave BackgroundToward fulfilling the requirements of PHY 210Srivas PrasadIn this report, I present the results of my investigations of the temperature of the cosmic microwave backgro
S.F. State - ECON - 312
THE CALIFORNIA TEST SCORE DATA SETThe California Standardized Testing and Reporting (STAR) dataset contains data on test performance, school characteristics and student demographic backgrounds. The data used here are from all 420 K-6 and K-8 districts in
S.F. State - ECON - 312
Documentation for CPS04 Data Each month the Bureau of Labor Statistics in the U.S. Department of Labor conducts the Current Population Survey (CPS), which provides data on labor force characteristics of the population, including the level of employment, u
Arkansas - CS - 2130
THE PRINCIPAL ELEMENTS OF THE NATURE OF SCIENCE: DISPELLING THE MYTHSWilliam F. McComas Rossier School of Education - WPH Univerisity of Southern California Los Angeles, CA 90089-0031 Adapted from the chapter in W. F. McComas (ed.) The Nature of Science
UCSC - LING - 201
Ideology, Power, and Linguistic TheoryGeoffrey K. PullumUniversity of California, Santa CruzMy aim in this paper is to discuss an intensely complex cluster of interlinked concepts involving distinctions between (i) descriptive and prescriptive grammar,
Oregon State University - ECE - 474
Oregon State University - ECE - 474
Oregon State University - ECE - 474
Oregon State University - ECE - 474
Oregon State University - ECE - 474
Verification of Large DesignsIncreasing system complexity is resulting in more chip respins due to functional errors.An ASIC respin costs \$200,000 to 2,000,000.To avoid respins verification must be more thorough even though chip complexity is skyrocket
Oregon State University - ECE - 474
University of Rochester - PHYS - 220
Name: Partners:PHYSICS 220 LAB #6: CIRCULAR M OTIONThe picture above is a copy of Copernicus drawing of the orbits of the planets which are nearly circular. It appeared in a book published in 1543. Since the direction of their motion is changing the pla
University of Rochester - PHYS - 107
Physics 107 Lecture 11: Exam 1 Review March 19th , 2002Today Wed. by 3pm Thursday Exam 1 (Ch 1,2,3,4,5,15) Review Exam 19:30 10:501HW8 Redo HW 10 All redosAdministrative: I will grade all Homework and Redos as quickly as possible and place them in a
Cornell - CS - 381
CS381 Fall 2002Final ExamThursday Dec 19, 2002 Location Olin 155 12:00-2:30pmThis is a 2 and hour in class closed book exam. All questions are straightforward and you should have no trouble doing them. Please show all work and write legibly. Thank you.
Cornell - CS - 381
CS381 Fall 2003Final ExamFriday Dec 12, 2003 Olin 255 9:00-11:30This is a 2 1 hour in class closed book exam. All questions are straightforward and you 2 should have no trouble doing them. Please show all work and write legibly. Thank you. 1. Is it dec
Cornell - CS - 381
CS381 Fall 2002Second Mid TermFriday Nov 8, 2002 Olin 155 9:05-9:55This is a 50-minute in class closed book exam. All questions are straightforward and you should have no trouble doing them. Please show all work and write legibly. Thank you. 1. (a) Giv
Cornell - CS - 381
CS381 Fall 2003Second Mid TermFriday Nov 7, 2003 Olin 255 9:05-9:55This is a 50-minute in class closed book exam. All questions are straightforward and you should have no trouble doing them. Please show all work and write legibly. Thank you. 1. Give a
Cornell - CS - 381
CS381 Fall 2003First Mid TermFriday Oct 10, 2003 Olin 255 9:05-9:55This is a 50-minute in class closed book exam. All questions are straightforward and you should have no trouble doing them. Please show all work and write legibly. Thank you. 1. Let L b
Cornell - CS - 381
Final Exam CS 381 Fall 2004 Wednesday Dec 15, 2004 Upson B17 9-11:30This is a closed book exam. All questions are straight forward. Please show all work and write legibly. Credit will be based on both correctness and clarity of answers. 1. Let R be a reg
Cornell - CS - 381
CS381 Fall 2004 Hollister B14Second Mid Term Friday Nov 5, 2004 9:05-9:55This is a 50-minute in class closed book exam. All questions are straightforward and you should have no trouble doing them. Please show all work and write legibly. Credit will be b
Cornell - CS - 381
CS 381 Fall 2005Final Exam Closed bookWednesday Dec 14 Philips 101 7-9:30pmAll problems should be straight forward. Partial credit will depend on clarity and conciseness of your answer. Please do not put down correct but irrelevant information. 1. Cons
Cornell - CS - 381
CS381 Fall 2005Second Mid Term Olin 155Friday Nov 4, 2005 9:05-9:55This is a 50-minute in class closed book exam. All questions are straightforward and you should have no trouble doing them. Please show all work and write legibly. Thank you. 1. Is the
Cornell - CS - 381
CS 381 Fall 2006Second Mid TermMonday, Nov. 13, 2006 Upson B17This is a 50-minute in class closed book exam. All questions are straightforward and you should have no trouble doing them. Please show all work and write legibly. Grading will take intoacc
Cornell - CS - 381
CS 381 Fall 2006First Mid TermFriday, Oct. 27, 2006 Upson B17This is a 50-minute in class closed book exam. All questions are straightforward and you should have no trouble doing them. Please show all work and write legibly. Grading will take into acco
Stanford - MATH - 136
d d h d r p p p rh ` xc n ` qpq8rTrGfkWyc2GThu`gpu`b2gxoGb@wu`eu e mc e p nc ec d ` d h d rh c dca` Gi`WyGfddpzGbcfw_Goy2GgpFui`bbhGkWyfc2sGBWihgdfe2b4Y s B|s rbbb4GIhr v tT s s c ` e d h d n e x d n d q c e x d d h d r h c d c a ` uyggD8rTsGkWyfc2iu`y2f
St. Mary MD - STATS - 0607
%!PS-Adobe-2.0 %Creator: dvipsk 5.58f Copyright 1986, 1994 Radical Eye Software %Title: matrices.dvi %Pages: 9 %PageOrder: Ascend %BoundingBox: 0 0 612 792 %EndComments %DVIPSCommandLine: dvips matrices %DVIPSParameters: dpi=600, compressed, comments remo
St. Mary MD - STATS - 0607
%!PSAdobe3.0 %Title: (\376\377) %Version: 1 2 %Creator: (\376\377) %CreationDate: (D:20000118123539) %For: (\376\377) %DocumentData: Clean7Bit %LanguageLevel: 2 %BoundingBox: 0 0 612 792 %Pages: 16 %DocumentProcessColors: (atend) %DocumentSuppliedResource
New Mexico - ECE - 446
DAQ E SeriesE Series User ManualE Series User ManualFebruary 2007 370503K-01Support Worldwide Technical Support and Product Information ni.com National Instruments Corporate Headquarters 11500 North Mopac Expressway Worldwide Offices Australia 1800 30
Cincinnati State - CE - 118
LM124/LM224/LM324/LM2902 Low Power Quad Operational AmplifiersAugust 2000LM124/LM224/LM324/LM2902 Low Power Quad Operational AmplifiersGeneral DescriptionThe LM124 series consists of four independent, high gain, internally frequency compensated operat
UPenn - CIS - 510
6.8. THE POST CORRESPONDENCE PROBLEM4236.8The Post Correspondence ProblemThe Post correspondence problem (due to Emil Post) is another undecidable problem that turns out to be a very helpful tool for proving problems in logic or in formal language the
University of Rochester - PHYS - 231
Phys. 231 Use correct notation for vectors and scalars.Exam 22005 1Name_ Read all problems carefully before attempting to solve them. Your work must be legible, and the organization must be clear. Correct answers without adequate explanation will be
Berkeley - STAT - 210
BickelDoksum. Mathematical Statistics. Volume I.July 2005 Errata p. xiii, l. 5. terrabytes should be terabytes. p. xiv, l. -20. Chapters 710 should be Chapters 79. p. 7, l. 13: Insert a well dened parameter after we can dene. p. 9 In the statement of (2)
Weber - PHSX - 2220
PHYS2220: Introduction to Physics IIThe Point:This course is a continuation of Physics I. We will begin with waves, then electricity and magnetism, and proceed through to quantum mechanics. You've now left the part of physics where your intuition about
Ball State - HIST - 415
Dr. Michael Wm. Doyle Department of History Burkhardt Bldg. 213 Ball State University Muncie, IN 47306-0480Office Phone: 765-285-8732; Fax: 765-285-5612 E-Mail: mwdoyle@bsu.edu Home Page: http:/www.bsu.edu/classes/doyle/hist415 Office Hrs.: Wed., 2:00-4:
University of Iowa - CS - 4020
Chapter 11 AXIOMATIC SEMANTICSThe techniques for operational semantics, introduced in Chapters 5 through 8, and denotational semantics, discussed in Chapters 9 and 10, are based on the notion of the &quot;state of a machine&quot;. For example, in the denotational
National Taiwan University - AEDE - 800
AED Economics 800 Research Methods Autumn 2006Final Examination Time Allowed: 2 hours Alan Randall December 5, 2006Answer questions 1, 2, and 3. For questions 1 and 2, logical rigor and correct use of concepts and definitions in your answer will enhance
University of Texas - ECE - 382
1 VLSI Design EXAM. II J. Abraham November 20, 2006 Name: Exam, No. 2 Open Book, Open Notes. Time Limit: 75 minutes (pace yourself ). Check for 6 pages in exam. Fall 2006Write all your answers in the spaces/boxes provided. Show any calculations in thes
University of Toronto - FOR - 3011
MFC Denmark and Sweden International Field Camp May 4 16, 2008 TRIP ITINERARY AND HOSTEL INFORMATION158 91614137 6 512ItinerarySunday 4 May Copenhagen Evening accommodation: all meet at Danhostel Copenhagen City by 8 pm. DANISH HOST: Professor Je
Athens Tech - EE - 126
Introduction to Probability: Problem Solutions(last updated: 5/15/07)cDimitri P. Bertsekas and John N. TsitsiklisMassachusetts Institute of TechnologyWWW site for book information and orders http:/www.athenasc.comAthena Scientic, Belmont, Massachuse
University of Toronto - FOR - 3011
Maps to the hostel booked for Sunday nights of 4 &amp; 11 May 2008 Danhostel Copenhagen City H.C. Andersens Boulevard 50, 1553 Kbenhavn V Tlf.: +45 3318 8332 or +45 3311 8585 Fax: +45 3311 8588 copenhagencity@danhostel.dk http:/www.danhostel.dk/copenhagencity
Oregon State - MEGA - 128
mega128.1-User GuideTekBotsTMOregon State UniversityVersion 1.2By Adriaan Smit OSU EE Graduate StudentCopyright 2003 - Oregon State UniversityTekBotsTMPage 2 of 32IndexBy Adriaan Smit . 1 OSU EE Graduate StudentIndex. 1 Index. 2 Introduction. 3
New Mexico - EECE - 541
wcfw_ucfw_po rn zu t w6 s ~pV|q w rnoppmnou w W Ccfw_ HIcfw_ w HP q q ycfw_ w p o w q n m u y cv w p o Hs qEe cfw_ p2w st B r u q oosEw q ycfw_ pu ttq o u sp Cv~ o ucfw_ s X z q 7 7ry qwHsw u Fs ~ srscfw_r
New Mexico - EECE - 541
Rayleigh and Rician FadingConsider two independent normal random variables X N (m1 , 2 ) and Y N (m2 , 2 ). Let us dene a complex Gaussian random variable Z via : Z = X + jY . The envelope and phase of the random variable Z are dened as: Y R = X 2 + Y 2
New Mexico - EECE - 541
%!PS-Adobe-2.0 %Creator: dvips 5.58 Copyright 1986, 1994 Radical Eye Software %Title: primer.dvi %Pages: 39 %PageOrder: Ascend %BoundingBox: 0 0 612 792 %EndComments %DVIPSCommandLine: dvips -o primer.ps primer.dvi %DVIPSParameters: dpi=300, comments remo
New Mexico - EECE - 541
%!PSAdobe3.0 %Title:(MO3934074007B1055.pdf) %Version:12 %CreationDate:(D:19971121071858) %DocumentData:Clean7Bit %LanguageLevel:2 %BoundingBox:00612792 %Pages:1 %DocumentProcessColors:(atend) %DocumentSuppliedResources: %+font:(atend) %+procset(AdobeAcrob
Caltech - ECE - 459
University Program Design Laboratory PackageUser GuideAugust 1997, ver. 1IntroductionThe University Program Design Laboratory Package was designed to meet the needs of universities teaching digital logic design with state-ofthe-art development tools
Pittsburgh - LEC - 2651
(Slide 5 in Lecture 8: Human Exposure Assessment II)Simplified Complexity of Human Exposure AssessmentAir Water Food Soil Surface Ingestion Dermal Inhalation Dietary
UCCS - CS - 526
SGFR: Secure Groupware for First Responders http:/cs.uccs.edu/~sgfr/ Contact: chow@cs.uccs.edu A NISSC Sponsored ProjectC. Edward Chow (PI) Chip Benight (PI) Ganesh GodavariDepartment of Computer SciencePart of this work is based on research sponsored
CUNY Baruch - CIS - 172
Chapter 1 Introduction to Computers and C+ ProgrammingOutline 1.1 Introduction 1.2 WhatisaComputer? 1.3 ComputerOrganization 1.4 EvolutionofOperatingSystems 1.5 PersonalComputing,DistributedComputingand Client/ServerComputing 1.6 MachineLanguages,Assembl
CUNY Baruch - WEB - 3300
SIGGRAPH 2001 Course 8An Introduction to the Kalman FilterGreg Welch Gary BishopUniversity of North Carolina at Chapel Hill Department of Computer Science Chapel Hill, NC 27599-3175 http:/www.cs.unc.edu/~cfw_welch, gb cfw_welch, gb@cs.unc.eduCopyright
CUNY Baruch - WEB - 7501
SIGGRAPH 2001 Course 8An Introduction to the Kalman FilterGreg Welch Gary BishopUniversity of North Carolina at Chapel Hill Department of Computer Science Chapel Hill, NC 27599-3175 http:/www.cs.unc.edu/~cfw_welch, gb cfw_welch, gb@cs.unc.eduCopyright