Parallel Simulation of Equation-Based Models on CUDA-Enabled GPUs

Size: px
Start display at page:

Download "Parallel Simulation of Equation-Based Models on CUDA-Enabled GPUs"

Transcription

1 Parallel Simulation of Equation-Based Models on CUDA-Enabled GPUs Per Ostlund Department of Computer and Information Science Linkoping University SE Linkoping, Sweden Kristian Stavaker Department of Computer and Information Science Linkoping University SE Linkoping, Sweden Peter Fritzson Department of Computer and Information Science Linkoping University SE Linkoping, Sweden ABSTRACT Our contributions wit tis work are metods and a prototype implementation for compiling and executing a limited set of equation-based matematical models (written in te object-oriented equation-based modeling language Modelica) on CUDA-enabled GPUs. We look at metods of finding parallelism in Modelica models, tat can be used on te massively parallel CUDA arcitecture. Te metods ave been implemented in a new back-end module of te OpenModelica compiler (an open-source Modelica compiler). Tis paper sows tat it is possible to automatically generate simulation code for pure continuous-time models tat can be reduced to an ordinary differential equation system witout algebraic loops and were te initial values of all variables and parameters are known at compile time. It is possible to get some speedup compared wit simulation on a single CPU core, a (approximated) relative speedup of 4.6 was for instance obtained for one model. Categories and Subject Descriptors D.3.4 [Programming Languages]: Processors Code generation General Terms Algoritms Keywords CUDA, GPU, GPGPU, Modelica, Matematical Simulation 1. INTRODUCTION Grapics Processing Units (GPUs) ave in recent years become increasingly programmable, giving rise to an emerging field of General-Purpose computation on Grapics Processing Units (GPGPU) [7]. Te teoretical processing power of GPUs ave far surpassed tat of CPUs due to te igly Permission to make digital or ard copies of all or part of tis work for personal or classroom use is granted witout fee provided tat copies are not made or distributed for profit or commercial advantage and tat copies bear tis notice and te full citation on te first page. To copy oterwise, to republis, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. POOSC OCT-2010, Reno, USA Copyrigt 2010 ACM /10/10...$ parallel structure of GPUs. [14] Harnessing te performance of GPUs, owever, is problematic. GPUs are in general good at solving certain massive parallel problems, for instance problems tat exists in te field of grapic computation were computations are applied pixel by pixel. In tis work, owever, we are dealing wit a more irregular problem, as we sall see. Compute Unified Device Arcitecture (CUDA) is a software platform for Nvidia GPUs tat simplifies GPGPU. Modelica is a language for equation-based object-oriented matematical modeling tat is being developed troug an international effort [6]. Compilation of Modelica models is a complex process tat involves many steps. Te end result of te compilation process is primarily a system of equations tat is linked wit run-time simulation code. Generation of parallel executable code from Modelica models as been a researc topic for several years at our researc group, see for instance [1], [5]. For oter work on parallel differential equation solver implementations, see for instance [10], [11], [4]. Our contributions wit tis work are metods and a prototype implementation for compiling and executing a limited set of equation-based matematical models, written in Modelica, on CUDA-enabled GPUs. Since Modelica is a large and complex language we restrict our models in tis work to pure continuous-time models (discrete time and ybrid models are also possible in Modelica) tat can be reduced to an explicit Ordinary Differential Equation (ODE) system witout algebraic loops and were te initial values of all variables and parameters are known at compile time. Te metods ave been implemented in a new back-end module of te OpenModelica compiler, an open-source Modelica compiler developed by te Open Source Modelica Consortium (OC) [9]. We obtained, for instance, a relative speedup of 4.6 for te WaveEquationSample model (presented in section 3) wit 3840 sections (about 7680 state variables), comparing te Nvidia GeForce 8800 GTS to te Intel E6600 CPU using single precision calculations. See section 5 for more details about tis measurement. A different and less general approac to parallelism tat as potential for better speedup for some models, for instance te WaveEquation- Sample model introduced in section 3, is to compile data parallel Modelica array computations into CUDA code. A first feasibility study of tis approac can be found in [12]. Tis paper is mainly based on [13] wic is an application of te work in [1] and [5].

2 Te rest of te paper is organized as follows. Section 2 contains background information on te languages and tools involved in tis work. Section 3 introduces one of te models we wis to generate code for and simulate. In section 4 we describe our implementation and in section 5 we provide measurements of generating and executing code wit our new implementation. Finally in section 6 we discuss our results and in section 7 we give some general conclusions and discuss future work. 2. BACKGROUND TECHNOLOGIES Tis section contains background information on te languages and systems involved in tis work. 2.1 Te Modelica Modeling Language Modelica is a language for equation-based object-oriented matematical modeling and it is being developed troug an international effort [6]. Modelica allows te user to write is/er equations directly in te model code witout aving to manually transform te equations into computational low-level form in a standard imperative programming language. Instead tis is done by te Modelica compiler in question. Te user can also make use of ig-level concepts suc as object-oriented programming and component composition. Te multi-domain capability of Modelica gives te user te possibility of combining electrical, mecanical, ydraulic, termodynamic, etc., model components witin te same application model. 2.2 OpenModelica OpenModelica is an open source implementation of a Modelica compiler, simulator and development environment for researc, education and industrial purposes. OpenModelica is developed and supported by an international effort, te Open Source Modelica Consortium (OC) [9]. It consists of a Modelica compiler as well as oter tools tat form an environment for creating and simulating Modelica models. 2.3 Compute Unified Device Arcitecture Streaming Multiprocessor Instruction Unit Data cace SFU SFU Sared memory Grapics Processing Unit Off-cip DRAM Figure 1: Simplified scematic of a GPU wit 96 s. For a more detailed description te reader is referred to [7]. A CUDA-enabled GPU consists of several Streaming Multiprocessors (s) wom eac contain several Scalar Processors (s). See Figure 1. In eac clock cycle, eac can produce a result. All te s in one execute te same instruction. Te s execute asyncronously witout communication and no formal consistency model exists between tem. One could say tat we ave SIMD execution (or SIMT execution, see [7]) inside one and MIMD execution across several s. Some syncronization between te s is possible via te global GPU memory. Te programmer launces blocks of treads from te ost (normally a CPU) to be executed on te GPU. Te blocks are automatically sceduled onto te different s. Eac block is divided into several warps (typically 32 treads). One warp is run at a time on eac and if one warp is stalled (due to memory latency, etc.) anoter warp may execute. If any treads in a warp take divergent execution pats, ten eac of tese pats will be executed separately, and te treads will ten converge again wen all pats ave been executed. Tere are several different types of memories on a CUDAenabled GPU. Eac as a set of registers. Access typically requires no extra clock cycles per instruction, except for some special cases. Eac as a sared memory (16 Kbytes on te Tesla C1060) wic is sared by all te s. Accessing te sared memory is typically as fast as a accessing a register and te sared memory can be used for cooperation between te different treads in a block. Eac also as a constant and a texture cace. We don t use te constant and texture caces in tis work. Finally we ave an off-cip Dynamic Random Access Memory (DRAM) (4 Gbytes on te Tesla C1060). Tis memory is accessible from all s as well as externally e.g. from te CPU wit DMA transfers. Te DRAM is muc slower for te treads to access tan te sared memory, typically it takes 400 to 600 clock cycles for a memory access. 3. CASE STUDY Here we introduce one of te models we wis to simulate in order to introduce te Modelica language and give te user an idea of te problem at and. Tis model is taken from [3](page 584). Te one-dimensional wave equation is given by a partial differential equation of te following form: 2 p t = 2 p 2 c2 x. (1) 2 were p = p(x, t) is a function of bot space and time. We consider a duct of lengt 10 were we let 5 x 5. We discretize te problem in te spatial dimension and approximate te spatial derivatives using difference approximations using te approximation: 2 p t = (p i 1 + p i+1 2p i) 2 c2 (2) x 2 were p i = p(x 1 + (i 1) x) on an equidistant grid and x is a small cange in distance. We get te following Modelica model, were te initial pressure is 0.0 and were n can be altered to increase/decrease te number of sections and tus te number of state variables:

3 model WaveEquationSample parameter Real L = 10 "Lengt of duct"; parameter Integer n = 30 "Number of sections"; parameter Real dl = L/n "Section lengt"; parameter Real c = 1; Real[n] p(start = fill (0,n)); Real[n] dp(start = fill (0,n)); equation p[1] = exp(-(-l/2)^2); p[n] = exp(-(l/2)^2); dp = der(p); for i in 2:n-1 loop der(dp[i])=c^2*(p[i+1]-2*p[i]+p[i-1])/dl^2; end for; end WaveEquationSample; Tis model consists of two sections, a section wit parameters and variables and a section wit equations. Parameters are constants tat can only cange teir values between simulation runs. Te vectors p and dp are vectors wit state variables since tese variables occur wit te derivative operator. Te for loop is a for-equation tat will be expanded into scalar equations. All of te equations in te model are merged togeter in one system and andled as described in section IMPLEMENTATION In tis section we start wit a sort overview of compilation and simulation of Modelica code and ten we discuss our new back-end module for te OpenModelica Compiler tat generates CUDA code. 4.1 Compilation and Simulation of Modelica Models Due to te special nature of Modelica, te compilation process of Modelica code differs quite a bit from imperative programming languages suc as C, C++, and Java. For a more detailed description te reader is referred to for instance [2], [3]. Te output from te OpenModelica front-end is an internal data structure wit separate lists for variables, equations, functions and algoritm sections. Te mapping of time-invariant parts of a model into executable code is done in a relatively straigtforward manner. Te andling of te equations is muc more complex toug and involves among oter tings symbolic index reduction, topological sorting (according to te causal dependencies between te equations), conversion into assignment form, etc. Simulation corresponds to solving te resulting equation system wit respect to time using a numerical solver metod. Te compiled Modelica model (te final equation system and te compiled Modelica functions) is linked wit a simulation runtime system tat contains among oter tings an integration metod. Fourt-order Runge-Kutta is te numerical integration metod used in tis paper. Initial values are given in a data file (tese ave been extracted from te original model) and a final plot file is produced (containing values for te state and algebraic variables at different time intervals). 4.2 CUDA Code Generation In [1] it was investigated ow parallel executable code from equation-based models could be generated, wic resulted in a OpenModelica back-end module. Tis work was ten Modelica Model Front-End, Equation Handling,... DAELow CUDA Codegen Task Grap Creation Task Grap Task Merging Sceduling Equations Code Generation CUDA Code Merged Task Grap Sceduled Tasks Figure 2: Te process of compiling a Modelica model to CUDA code. continued in [5] by inlining te numerical solver and by introducing so-called software pipelining. In tis work we ave tried to adopt some of tese tecniques, especially from [1], for generating CUDA code. Te reader is referred to [13] for a more in-dept description. See Figure 2 for te different compilation pases Task Grap Creation and Task Merging Te CUDA code generation module is invoked in te backend of te compiler wit a sorted equation system as input. As a first step a task grap is generated from te equation system. A task grap is a directed acyclic grap were eac vertex represents a task (a scalar operation) and te edges of te grap represents precedence constraints on te tasks. Tere is a cost associated wit eac edge and task. Te cost of an edge is te cost of sending data between two tasks and te cost of a task is te cost of executing te task. So far we ave only used crude costs for te tasks: te cost of unary and binary operations are set to 1 and te cost of special functions are set to 4 (tis sould reflect te fact tat a as eigt s but only two special function units). Te cost of communication is set to 100, wic sould reflect te latencies introduced by te global memory. A task grap is a convenient structure to work wit for detecting parts of an ODE system tat can be computed in parallel. Because te initial task grap is extremely fine grained, a merging algoritm is applied to it. Scalar tasks are merged togeter into larger nodes Sceduling Wen te task grap as been merged it is necessary to determine in wic order te tasks sould be executed, and weter tey sould be executed in parallel on different s or not. Te sceduling is done in two main steps. In te first step te nodes in te merged task grap is sceduled wit te so-called critical pat algoritm and ten te tasks

4 inside eac node are sceduled. Te critical pat algoritm selects te critical pat in a grap, wic is te pat wit te longest execution time. Te critical pat algoritm is run over and over again until te wole grap as been covered. Te tasks inside one node are sceduled using a first in, first out queue wit te tasks to be sceduled. An example can be seen in figure 3. a c d e f b g g e c f d b a Time Figure 3: An example of te task sceduling algoritm. In addition to tese two steps te sceduler also tries to find nodes tat are operationally equivalent to oter nodes, and scedules tose nodes to be executed in parallel on te same (SIMD style execution). As of now we use a relatively crude test for operational equality, see [13] for more details. Execution Pat Execution Pat List Processor Scedule Figure 4: An example of a scedule for two streaming multiprocessors. Te result of te sceduling is a processor scedule. See figure 4. Te processor scedule contains execution pats and execution pat lists. An execution pat is a list of task executed in order. One executes one execution pat list, tat opefully contains several execution pats, at a time. Inside one we can, if tere are several execution pats, execute in SIMD mode. Running several s in parallel and syncronizing via te global memory, corresponds to MIMD execution. Te blocks of treads are sceduled automatically on te different s and we do not know in wic order te different blocks are going to execute. However, we never execute more blocks tan tere are s (to avoid dead-locks) and we syncronize via te global memory, tus we can run several blocks in parallel. Communication between processors is necassary in many cases. Eac node as to be inspected to see if any of te task as a dependency tat is sceduled on anoter processor. If tis is te case te sceduler inserts signals and locks into te scedule and determines wat data sould be sent were. Special execution pats for communication are inserted into te scedule Code Generation Code generation is done by iterating troug te scedule one processor at a time and one execution pat list at a time. If te execution pat list is a special execution pat for communication, communication code is generated. If te execution pat list contains several execution pats code for SIMD is generated. Oterwise if it is a normal execution pat list, code for eac task is generated one by one. A tecnique to reduce long off-cip DRAM latencies is memory coalescing. Te DRAM memory can access wole cunks of memory at a time. We use coalesced memory reads tat read 16 variables at a time from te device memory, te size of a coalesced read of 32-bit variables, and ten we move tose variables to were tey sould be in te sared memory on an. It does not matter if not all 16 variables are used since it costs as muc to read one variable as it does to read 16 variables. By coalescing memory accesses and aving several warps active at te same time it is tus possible to mask te latencies associated wit te off-cip memory somewat Generated Code In te code listning below te main simulation function is given. A fourt order Runge-kutta integration sceme is used. Tis integration sceme was used bot for te GPU implementation and te normal sequential CPU implementation, tat we used for comparison. Tis integration metod is relatively easy to implement yet gives decent enoug results. k 1 = f (x (t), u (t), t) k 2 = f ( x ( t + ) ( ) ) 2 k1, u t + 2, t + 2 k 3 = f ( x ( t + ) ( ) ) 2 k2, u t + 2, t + 2 k 4 = f (x (t + k 3), u (t + ), t + ) x(t + ) x(t) + 1 (k1 + 2 k2 + 2 k3 + k4) 6 Te first four steps in te metod evaluates f at different points in time, and in te last step a weigted sum of tese values is calculated. Te next value of x is ten calculated using te weigted sum. Te execution of tasks and incrementation of steps are done in parallel by launcing kernels. //Determine te size of te sared memory needed. int smem_size = 100 * sizeof(real); for(int step = 0; step < steps; ++step) { //Move te pointers of te result arrays forward. r_dx += DERIVATIVES; r_x += STATES; r_y += ALGEBRAICS; //Execute te tasks, call integration kernel. execute_tasks<<<7, 20, smem_size>>>(d_dx, d_x, step_and_increment1<<<2, 32>>>(d_x, d_old_x, d_dx, d_k, alf_); //Increment te time by alf a time step.

5 GeForce Tesla 8800 C1060 GTS Streaming Multiprocessors Scalar Processors Scalar Processor Clock (MHz) Single Precision GFLOPS Double Precision GFLOPS N/A 78 Memory Amount (MB) Memory Interface 320-bit 512-bit Memory Clock (MHz) Memory Bandwidt (GB/s) PCIe Version (1.0 used) PCIe Bandwidt (GB/s) 4 8 (4 used) CUDA Compute Capability Table 1: GPU Specifications Execution Time (seconds) E GTS C1060 single C1060 double N t += alf_; //Do two more steps of te RK4 metod. execute_tasks<<<7, 20, smem_size>>>(d_dx, d_x, step_and_increment2<<<2, 32>>>(d_x, d_old_x, d_dx, d_k, alf_); execute_tasks<<<7, 20, smem_size>>>(d_dx, d_x, step_and_increment3<<<2, 32>>>(d_x, d_old_x, d_dx, d_k, ); //Increment te time again wit alf a time step. t += alf_; //Do te final integration. execute_tasks<<<7, 20, smem_size>>>(d_dx, d_x, step_and_integrate<<<2, 32>>>(d_x, d_old_x, d_dx, d_k, _div_6); //Save te new values. cudamemcpy(r_x, d_x, STATES * sizeof(real), cudamemcpydevicetohost); cudamemcpy(r_dx, d_dx, DERIVATIVES * sizeof(real), cudamemcpydevicetohost); cudamemcpy(r_y, d_y, ALGEBRAICS * sizeof(real), cudamemcpydevicetohost); } 5. EXPERIMENTAL RESULTS In tis section we present measurements of executing CUDA code from one model on two GPU arcitectures and on one CPU arcitecture. For more measurement data see [13]. Table 1 sows te specifications of te two GPU arcitectures used. Figure 5 sows te time-measurements obtained. We alter te parameter n for eac simulation run. Te CPU used was an Intel Core 2 Duo E6600 (2.4 GHz clock frequency) but only one core was used (since tere were no code generator tat generates parallel code from OpenModelica available). Te model used, WaveEquationSample, was introduced in section 3 and section 6 contains a discussion of te results. An approximation was done tat gives an upper limit of te performance, see [13]. Figure 5: Execution Time for te WaveEquation- Sample Model as a Function of te Number of Sections GTS C1060 C1060 single double Task Execution Sared Mem Allocation Integration Memory Transfers Table 2: Seconds spent in te different parts of te simulation of te WaveEquationSample model. 6. DISCUSSION OF RESULTS We obtained a relative speedup of 4.6 wit n=3840 for WaveEquationsSample (comparing te GeForce 8800 GTS to te Intel E6600 CPU using single precision calculations). As noted before, tis was an approximation. For small models (few state variables) te GPU performs badly but as te number increases te GPU retain a relatively flat curve, wile te simulation times on te CPU rises muc faster. Te reason for tis is likely tat te GPUs need many tread blocks wit many treads to fully utilize teir power, and smaller models terefore do not use all s available. Te simulation times on te CPUs instead beaves as could be expected by approximately doubling wen te model size is doubled. It was sown tat memory transactions take most of te time, and te actual computations take a smaller amount of time. See table 2. Wile te models used in tis work are parallel enoug to easily spawn many treads, te computations per variable ratio is quite low. Models wit more computations per variable sould see an even larger performance increase wen using a GPU.

6 7. CONCLUSIONS Te goal of tis work was to evaluate te feasibility of simulating Modelica models on CUDA-enabled GPUs. We ave demonstrated tat it is possible to simulate at least a subset of typical models. Nvidia will soon release te new Fermi arcitecture [8]. Tis arcitecture as several improvements. It as a cace ierarcy, improved support for double precision operations, more sared memory on te s, support for running several kernels at a time (better utilization of idle s), etc. A potential future work could terefore be to compile CUDA code for te Fermi arcitecture using te metods outlined in tis paper. A different approac of using GPUs and CUDA for simulating Modelica models (or a subset of Modelica models) is to compile Modelica array-equations into CUDA code. Modelica models containing for-equations operating over large arrays are for instance common in models derived from discretizations of partial differential equations. Tis is recently initialized work at our researc group, see [12]. 8. ACKNOWLEDGMENTS Funded by te European ITEA2 OPENPROD Project (Open Model-Driven Wole-Product Development and Simulation Environment) and CUGS Graduate Scool. 9. REFERENCES [1] P. Aronsson. Automatic Parallelization of Equation-Based Simulation Programs. PD tesis, Linköping University, Sweden, Dissertation No [2] E. K. Francois Cellier. Continuous System Simulation. Springer, [3] P. Fritzson. Principles of Object-Oriented Modeling and Simulation wit Modelica 2.1. IEEE Press, [4] M. Korc and T. Rauber. Scalable parallel rk solvers for odes derived by te metod of lines. In H. Kosc, L. Böszörményi, and H. Hellwagner, editors, Euro-Par, volume 2790 of Lecture Notes in Computer Science, pages Springer, [5] H. Lundvall. Automatic Parallelization using Pipelining for Equation-Based Simulation Languages, Licentiate tesis Linköping University, [6] Modelica and te Modelica Association, accessed January ttp:// [7] CUDA Zone, accessed January ttp:// ome new.tml. [8] Nvidia Fermi, accessed January ttp:// arcitecture.tml. [9] OpenModelica, accessed January ttp:// [10] T. Rauber and G. Rünger. Iterated runge-kutta metods on distributed memory multiprocessors. In PDP, pages IEEE Computer Society, [11] T. Rauber and G. Rünger. Parallel execution of embedded and iterated runge-kutta metods. Concurrency - Practice and Experience, 11(7): , [12] K. Stavåker, D. Rolls, J. Guo, P. Fritzson, S. Scolz. Compilation of Modelica Array Computations into Single Assignment C for Efficient Execution on CUDA-enabled GPUs. EOOLT 2010, Oslo, Norway, October 3, [13] P. Östlund. Simulation of Modelica Models on te CUDA Arcitecture. Master Tesis. LIU-IDA/LITH-EX-A 09/062 SE. Linköping University, [14] M. Wolfe, Te Portland Group, Inc: Compilers and More: GPU Arcitecture and Applications, September

Bounding Tree Cover Number and Positive Semidefinite Zero Forcing Number

Bounding Tree Cover Number and Positive Semidefinite Zero Forcing Number Bounding Tree Cover Number and Positive Semidefinite Zero Forcing Number Sofia Burille Mentor: Micael Natanson September 15, 2014 Abstract Given a grap, G, wit a set of vertices, v, and edges, various

More information

2 The Derivative. 2.0 Introduction to Derivatives. Slopes of Tangent Lines: Graphically

2 The Derivative. 2.0 Introduction to Derivatives. Slopes of Tangent Lines: Graphically 2 Te Derivative Te two previous capters ave laid te foundation for te study of calculus. Tey provided a review of some material you will need and started to empasize te various ways we will view and use

More information

A Cost Model for Distributed Shared Memory. Using Competitive Update. Jai-Hoon Kim Nitin H. Vaidya. Department of Computer Science

A Cost Model for Distributed Shared Memory. Using Competitive Update. Jai-Hoon Kim Nitin H. Vaidya. Department of Computer Science A Cost Model for Distributed Sared Memory Using Competitive Update Jai-Hoon Kim Nitin H. Vaidya Department of Computer Science Texas A&M University College Station, Texas, 77843-3112, USA E-mail: fjkim,vaidyag@cs.tamu.edu

More information

4.1 Tangent Lines. y 2 y 1 = y 2 y 1

4.1 Tangent Lines. y 2 y 1 = y 2 y 1 41 Tangent Lines Introduction Recall tat te slope of a line tells us ow fast te line rises or falls Given distinct points (x 1, y 1 ) and (x 2, y 2 ), te slope of te line troug tese two points is cange

More information

Computer Physics Communications. Multi-GPU acceleration of direct pore-scale modeling of fluid flow in natural porous media

Computer Physics Communications. Multi-GPU acceleration of direct pore-scale modeling of fluid flow in natural porous media Computer Pysics Communications 183 (2012) 1890 1898 Contents lists available at SciVerse ScienceDirect Computer Pysics Communications ournal omepage: www.elsevier.com/locate/cpc Multi-GPU acceleration

More information

Section 2.3: Calculating Limits using the Limit Laws

Section 2.3: Calculating Limits using the Limit Laws Section 2.3: Calculating Limits using te Limit Laws In previous sections, we used graps and numerics to approimate te value of a it if it eists. Te problem wit tis owever is tat it does not always give

More information

Density Estimation Over Data Stream

Density Estimation Over Data Stream Density Estimation Over Data Stream Aoying Zou Dept. of Computer Science, Fudan University 22 Handan Rd. Sangai, 2433, P.R. Cina ayzou@fudan.edu.cn Ziyuan Cai Dept. of Computer Science, Fudan University

More information

Asynchronous Power Flow on Graphic Processing Units

Asynchronous Power Flow on Graphic Processing Units 1 Asyncronous Power Flow on Grapic Processing Units Manuel Marin, Student Member, IEEE, David Defour, and Federico Milano, Senior Member, IEEE Abstract Asyncronous iterations can be used to implement fixed-point

More information

2.8 The derivative as a function

2.8 The derivative as a function CHAPTER 2. LIMITS 56 2.8 Te derivative as a function Definition. Te derivative of f(x) istefunction f (x) defined as follows f f(x + ) f(x) (x). 0 Note: tis differs from te definition in section 2.7 in

More information

More on Functions and Their Graphs

More on Functions and Their Graphs More on Functions and Teir Graps Difference Quotient ( + ) ( ) f a f a is known as te difference quotient and is used exclusively wit functions. Te objective to keep in mind is to factor te appearing in

More information

Implementation of Digital Signal Processing Algorithm in General Purpose Graphics Processing Unit (GPGPU)

Implementation of Digital Signal Processing Algorithm in General Purpose Graphics Processing Unit (GPGPU) Implementation of Digital Signal Processing Algoritm in General Purpose Grapics Processing Unit (GPG) Jagdamb Beari Srivastava 1, R.K. pandey, Jitendra Jain 3 Programme Asst. (Computer), Jawaarlal Neru

More information

Asynchronous Power Flow on Graphic Processing Units

Asynchronous Power Flow on Graphic Processing Units Asyncronous Power Flow on Grapic Processing Units Manuel Marin, David Defour, Federico Milano To cite tis version: Manuel Marin, David Defour, Federico Milano. Asyncronous Power Flow on Grapic Processing

More information

4.2 The Derivative. f(x + h) f(x) lim

4.2 The Derivative. f(x + h) f(x) lim 4.2 Te Derivative Introduction In te previous section, it was sown tat if a function f as a nonvertical tangent line at a point (x, f(x)), ten its slope is given by te it f(x + ) f(x). (*) Tis is potentially

More information

MATH 5a Spring 2018 READING ASSIGNMENTS FOR CHAPTER 2

MATH 5a Spring 2018 READING ASSIGNMENTS FOR CHAPTER 2 MATH 5a Spring 2018 READING ASSIGNMENTS FOR CHAPTER 2 Note: Tere will be a very sort online reading quiz (WebWork) on eac reading assignment due one our before class on its due date. Due dates can be found

More information

An Algorithm for Loopless Deflection in Photonic Packet-Switched Networks

An Algorithm for Loopless Deflection in Photonic Packet-Switched Networks An Algoritm for Loopless Deflection in Potonic Packet-Switced Networks Jason P. Jue Center for Advanced Telecommunications Systems and Services Te University of Texas at Dallas Ricardson, TX 75083-0688

More information

Fast Calculation of Thermodynamic Properties of Water and Steam in Process Modelling using Spline Interpolation

Fast Calculation of Thermodynamic Properties of Water and Steam in Process Modelling using Spline Interpolation P R E P R N T CPWS XV Berlin, September 8, 008 Fast Calculation of Termodynamic Properties of Water and Steam in Process Modelling using Spline nterpolation Mattias Kunick a, Hans-Joacim Kretzscmar a,

More information

Numerical Derivatives

Numerical Derivatives Lab 15 Numerical Derivatives Lab Objective: Understand and implement finite difference approximations of te derivative in single and multiple dimensions. Evaluate te accuracy of tese approximations. Ten

More information

13.5 DIRECTIONAL DERIVATIVES and the GRADIENT VECTOR

13.5 DIRECTIONAL DERIVATIVES and the GRADIENT VECTOR 13.5 Directional Derivatives and te Gradient Vector Contemporary Calculus 1 13.5 DIRECTIONAL DERIVATIVES and te GRADIENT VECTOR Directional Derivatives In Section 13.3 te partial derivatives f x and f

More information

Hash-Based Indexes. Chapter 11. Comp 521 Files and Databases Spring

Hash-Based Indexes. Chapter 11. Comp 521 Files and Databases Spring Has-Based Indexes Capter 11 Comp 521 Files and Databases Spring 2010 1 Introduction As for any index, 3 alternatives for data entries k*: Data record wit key value k

More information

Symmetric Tree Replication Protocol for Efficient Distributed Storage System*

Symmetric Tree Replication Protocol for Efficient Distributed Storage System* ymmetric Tree Replication Protocol for Efficient Distributed torage ystem* ung Cune Coi 1, Hee Yong Youn 1, and Joong up Coi 2 1 cool of Information and Communications Engineering ungkyunkwan University

More information

Haar Transform CS 430 Denbigh Starkey

Haar Transform CS 430 Denbigh Starkey Haar Transform CS Denbig Starkey. Background. Computing te transform. Restoring te original image from te transform 7. Producing te transform matrix 8 5. Using Haar for lossless compression 6. Using Haar

More information

The Euler and trapezoidal stencils to solve d d x y x = f x, y x

The Euler and trapezoidal stencils to solve d d x y x = f x, y x restart; Te Euler and trapezoidal stencils to solve d d x y x = y x Te purpose of tis workseet is to derive te tree simplest numerical stencils to solve te first order d equation y x d x = y x, and study

More information

CHAPTER 7: TRANSCENDENTAL FUNCTIONS

CHAPTER 7: TRANSCENDENTAL FUNCTIONS 7.0 Introduction and One to one Functions Contemporary Calculus 1 CHAPTER 7: TRANSCENDENTAL FUNCTIONS Introduction In te previous capters we saw ow to calculate and use te derivatives and integrals of

More information

3.6 Directional Derivatives and the Gradient Vector

3.6 Directional Derivatives and the Gradient Vector 288 CHAPTER 3. FUNCTIONS OF SEVERAL VARIABLES 3.6 Directional Derivatives and te Gradient Vector 3.6.1 Functions of two Variables Directional Derivatives Let us first quickly review, one more time, te

More information

Hash-Based Indexes. Chapter 11. Comp 521 Files and Databases Fall

Hash-Based Indexes. Chapter 11. Comp 521 Files and Databases Fall Has-Based Indexes Capter 11 Comp 521 Files and Databases Fall 2012 1 Introduction Hasing maps a searc key directly to te pid of te containing page/page-overflow cain Doesn t require intermediate page fetces

More information

THANK YOU FOR YOUR PURCHASE!

THANK YOU FOR YOUR PURCHASE! THANK YOU FOR YOUR PURCHASE! Te resources included in tis purcase were designed and created by me. I ope tat you find tis resource elpful in your classroom. Please feel free to contact me wit any questions

More information

Multi-Stack Boundary Labeling Problems

Multi-Stack Boundary Labeling Problems Multi-Stack Boundary Labeling Problems Micael A. Bekos 1, Micael Kaufmann 2, Katerina Potika 1 Antonios Symvonis 1 1 National Tecnical University of Atens, Scool of Applied Matematical & Pysical Sciences,

More information

12.2 Techniques for Evaluating Limits

12.2 Techniques for Evaluating Limits 335_qd /4/5 :5 PM Page 863 Section Tecniques for Evaluating Limits 863 Tecniques for Evaluating Limits Wat ou sould learn Use te dividing out tecnique to evaluate its of functions Use te rationalizing

More information

Computing geodesic paths on manifolds

Computing geodesic paths on manifolds Proc. Natl. Acad. Sci. USA Vol. 95, pp. 8431 8435, July 1998 Applied Matematics Computing geodesic pats on manifolds R. Kimmel* and J. A. Setian Department of Matematics and Lawrence Berkeley National

More information

A UPnP-based Decentralized Service Discovery Improved Algorithm

A UPnP-based Decentralized Service Discovery Improved Algorithm Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol.1, No.1, Marc 2013, pp. 21~26 ISSN: 2089-3272 21 A UPnP-based Decentralized Service Discovery Improved Algoritm Yu Si-cai*, Wu Yan-zi,

More information

1.4 RATIONAL EXPRESSIONS

1.4 RATIONAL EXPRESSIONS 6 CHAPTER Fundamentals.4 RATIONAL EXPRESSIONS Te Domain of an Algebraic Epression Simplifying Rational Epressions Multiplying and Dividing Rational Epressions Adding and Subtracting Rational Epressions

More information

An Anchor Chain Scheme for IP Mobility Management

An Anchor Chain Scheme for IP Mobility Management An Ancor Cain Sceme for IP Mobility Management Yigal Bejerano and Israel Cidon Department of Electrical Engineering Tecnion - Israel Institute of Tecnology Haifa 32000, Israel E-mail: bej@tx.tecnion.ac.il.

More information

Linear Interpolating Splines

Linear Interpolating Splines Jim Lambers MAT 772 Fall Semester 2010-11 Lecture 17 Notes Tese notes correspond to Sections 112, 11, and 114 in te text Linear Interpolating Splines We ave seen tat ig-degree polynomial interpolation

More information

Materials: Whiteboard, TI-Nspire classroom set, quadratic tangents program, and a computer projector.

Materials: Whiteboard, TI-Nspire classroom set, quadratic tangents program, and a computer projector. Adam Clinc Lesson: Deriving te Derivative Grade Level: 12 t grade, Calculus I class Materials: Witeboard, TI-Nspire classroom set, quadratic tangents program, and a computer projector. Goals/Objectives:

More information

CESILA: Communication Circle External Square Intersection-Based WSN Localization Algorithm

CESILA: Communication Circle External Square Intersection-Based WSN Localization Algorithm Sensors & Transducers 2013 by IFSA ttp://www.sensorsportal.com CESILA: Communication Circle External Square Intersection-Based WSN Localization Algoritm Sun Hongyu, Fang Ziyi, Qu Guannan College of Computer

More information

NOTES: A quick overview of 2-D geometry

NOTES: A quick overview of 2-D geometry NOTES: A quick overview of 2-D geometry Wat is 2-D geometry? Also called plane geometry, it s te geometry tat deals wit two dimensional sapes flat tings tat ave lengt and widt, suc as a piece of paper.

More information

Redundancy Awareness in SQL Queries

Redundancy Awareness in SQL Queries Redundancy Awareness in QL Queries Bin ao and Antonio Badia omputer Engineering and omputer cience Department University of Louisville bin.cao,abadia @louisville.edu Abstract In tis paper, we study QL

More information

Minimizing Memory Access By Improving Register Usage Through High-level Transformations

Minimizing Memory Access By Improving Register Usage Through High-level Transformations Minimizing Memory Access By Improving Register Usage Troug Hig-level Transformations San Li Scool of Computer Engineering anyang Tecnological University anyang Avenue, SIGAPORE 639798 Email: p144102711@ntu.edu.sg

More information

On the Use of Radio Resource Tests in Wireless ad hoc Networks

On the Use of Radio Resource Tests in Wireless ad hoc Networks Tecnical Report RT/29/2009 On te Use of Radio Resource Tests in Wireless ad oc Networks Diogo Mónica diogo.monica@gsd.inesc-id.pt João Leitão jleitao@gsd.inesc-id.pt Luis Rodrigues ler@ist.utl.pt Carlos

More information

Real-Time Wireless Routing for Industrial Internet of Things

Real-Time Wireless Routing for Industrial Internet of Things Real-Time Wireless Routing for Industrial Internet of Tings Cengjie Wu, Dolvara Gunatilaka, Mo Sa, Cenyang Lu Cyber-Pysical Systems Laboratory, Wasington University in St. Louis Department of Computer

More information

Optimal In-Network Packet Aggregation Policy for Maximum Information Freshness

Optimal In-Network Packet Aggregation Policy for Maximum Information Freshness 1 Optimal In-etwork Packet Aggregation Policy for Maimum Information Fresness Alper Sinan Akyurek, Tajana Simunic Rosing Electrical and Computer Engineering, University of California, San Diego aakyurek@ucsd.edu,

More information

PYRAMID FILTERS BASED ON BILINEAR INTERPOLATION

PYRAMID FILTERS BASED ON BILINEAR INTERPOLATION PYRAMID FILTERS BASED ON BILINEAR INTERPOLATION Martin Kraus Computer Grapics and Visualization Group, Tecnisce Universität Müncen, Germany krausma@in.tum.de Magnus Strengert Visualization and Interactive

More information

Our Calibrated Model has No Predictive Value: An Example from the Petroleum Industry

Our Calibrated Model has No Predictive Value: An Example from the Petroleum Industry Our Calibrated Model as No Predictive Value: An Example from te Petroleum Industry J.N. Carter a, P.J. Ballester a, Z. Tavassoli a and P.R. King a a Department of Eart Sciences and Engineering, Imperial

More information

UUV DEPTH MEASUREMENT USING CAMERA IMAGES

UUV DEPTH MEASUREMENT USING CAMERA IMAGES ABCM Symposium Series in Mecatronics - Vol. 3 - pp.292-299 Copyrigt c 2008 by ABCM UUV DEPTH MEASUREMENT USING CAMERA IMAGES Rogerio Yugo Takimoto Graduate Scool of Engineering Yokoama National University

More information

Notes: Dimensional Analysis / Conversions

Notes: Dimensional Analysis / Conversions Wat is a unit system? A unit system is a metod of taking a measurement. Simple as tat. We ave units for distance, time, temperature, pressure, energy, mass, and many more. Wy is it important to ave a standard?

More information

Integrating Multimedia Applications in Hard Real-Time Systems

Integrating Multimedia Applications in Hard Real-Time Systems Integrating Multimedia Applications in Hard Real-Time Systems Luca Abeni and Giorgio Buttazzo Scuola Superiore S. Anna, Pisa luca@arti.sssup.it, giorgio@sssup.it Abstract Tis paper focuses on te problem

More information

Vector Processing Contours

Vector Processing Contours Vector Processing Contours Andrey Kirsanov Department of Automation and Control Processes MAMI Moscow State Tecnical University Moscow, Russia AndKirsanov@yandex.ru A.Vavilin and K-H. Jo Department of

More information

Mean Shifting Gradient Vector Flow: An Improved External Force Field for Active Surfaces in Widefield Microscopy.

Mean Shifting Gradient Vector Flow: An Improved External Force Field for Active Surfaces in Widefield Microscopy. Mean Sifting Gradient Vector Flow: An Improved External Force Field for Active Surfaces in Widefield Microscopy. Margret Keuper Cair of Pattern Recognition and Image Processing Computer Science Department

More information

You Try: A. Dilate the following figure using a scale factor of 2 with center of dilation at the origin.

You Try: A. Dilate the following figure using a scale factor of 2 with center of dilation at the origin. 1 G.SRT.1-Some Tings To Know Dilations affect te size of te pre-image. Te pre-image will enlarge or reduce by te ratio given by te scale factor. A dilation wit a scale factor of 1> x >1enlarges it. A dilation

More information

Piecewise Polynomial Interpolation, cont d

Piecewise Polynomial Interpolation, cont d Jim Lambers MAT 460/560 Fall Semester 2009-0 Lecture 2 Notes Tese notes correspond to Section 4 in te text Piecewise Polynomial Interpolation, cont d Constructing Cubic Splines, cont d Having determined

More information

ANTENNA SPHERICAL COORDINATE SYSTEMS AND THEIR APPLICATION IN COMBINING RESULTS FROM DIFFERENT ANTENNA ORIENTATIONS

ANTENNA SPHERICAL COORDINATE SYSTEMS AND THEIR APPLICATION IN COMBINING RESULTS FROM DIFFERENT ANTENNA ORIENTATIONS NTNN SPHRICL COORDINT SSTMS ND THIR PPLICTION IN COMBINING RSULTS FROM DIFFRNT NTNN ORINTTIONS llen C. Newell, Greg Hindman Nearfield Systems Incorporated 133. 223 rd St. Bldg. 524 Carson, C 9745 US BSTRCT

More information

Announcements SORTING. Prelim 1. Announcements. A3 Comments 9/26/17. This semester s event is on Saturday, November 4 Apply to be a teacher!

Announcements SORTING. Prelim 1. Announcements. A3 Comments 9/26/17. This semester s event is on Saturday, November 4 Apply to be a teacher! Announcements 2 "Organizing is wat you do efore you do someting, so tat wen you do it, it is not all mixed up." ~ A. A. Milne SORTING Lecture 11 CS2110 Fall 2017 is a program wit a teac anyting, learn

More information

Chapter K. Geometric Optics. Blinn College - Physics Terry Honan

Chapter K. Geometric Optics. Blinn College - Physics Terry Honan Capter K Geometric Optics Blinn College - Pysics 2426 - Terry Honan K. - Properties of Ligt Te Speed of Ligt Te speed of ligt in a vacuum is approximately c > 3.0µ0 8 mês. Because of its most fundamental

More information

Intra- and Inter-Session Network Coding in Wireless Networks

Intra- and Inter-Session Network Coding in Wireless Networks Intra- and Inter-Session Network Coding in Wireless Networks Hulya Seferoglu, Member, IEEE, Atina Markopoulou, Member, IEEE, K K Ramakrisnan, Fellow, IEEE arxiv:857v [csni] 3 Feb Abstract In tis paper,

More information

Tuning MAX MIN Ant System with off-line and on-line methods

Tuning MAX MIN Ant System with off-line and on-line methods Université Libre de Bruxelles Institut de Recerces Interdisciplinaires et de Développements en Intelligence Artificielle Tuning MAX MIN Ant System wit off-line and on-line metods Paola Pellegrini, Tomas

More information

12.2 TECHNIQUES FOR EVALUATING LIMITS

12.2 TECHNIQUES FOR EVALUATING LIMITS Section Tecniques for Evaluating Limits 86 TECHNIQUES FOR EVALUATING LIMITS Wat ou sould learn Use te dividing out tecnique to evaluate its of functions Use te rationalizing tecnique to evaluate its of

More information

Multi-Objective Particle Swarm Optimizers: A Survey of the State-of-the-Art

Multi-Objective Particle Swarm Optimizers: A Survey of the State-of-the-Art Multi-Objective Particle Swarm Optimizers: A Survey of te State-of-te-Art Margarita Reyes-Sierra and Carlos A. Coello Coello CINVESTAV-IPN (Evolutionary Computation Group) Electrical Engineering Department,

More information

Section 1.2 The Slope of a Tangent

Section 1.2 The Slope of a Tangent Section 1.2 Te Slope of a Tangent You are familiar wit te concept of a tangent to a curve. Wat geometric interpretation can be given to a tangent to te grap of a function at a point? A tangent is te straigt

More information

Distributed and Optimal Rate Allocation in Application-Layer Multicast

Distributed and Optimal Rate Allocation in Application-Layer Multicast Distributed and Optimal Rate Allocation in Application-Layer Multicast Jinyao Yan, Martin May, Bernard Plattner, Wolfgang Mülbauer Computer Engineering and Networks Laboratory, ETH Zuric, CH-8092, Switzerland

More information

Investigating an automated method for the sensitivity analysis of functions

Investigating an automated method for the sensitivity analysis of functions Investigating an automated metod for te sensitivity analysis of functions Sibel EKER s.eker@student.tudelft.nl Jill SLINGER j..slinger@tudelft.nl Delft University of Tecnology 2628 BX, Delft, te Neterlands

More information

An experimental framework to investigate context-aware schemes for content delivery

An experimental framework to investigate context-aware schemes for content delivery An experimental framework to investigate context-aware scemes for content delivery Pietro Lungaro +, Cristobal Viedma +, Zary Segall + and Pavan Kumar + Mobile Service Lab, Royal Institute of Tecnology

More information

Fault Localization Using Tarantula

Fault Localization Using Tarantula Class 20 Fault localization (cont d) Test-data generation Exam review: Nov 3, after class to :30 Responsible for all material up troug Nov 3 (troug test-data generation) Send questions beforeand so all

More information

Coarticulation: An Approach for Generating Concurrent Plans in Markov Decision Processes

Coarticulation: An Approach for Generating Concurrent Plans in Markov Decision Processes Coarticulation: An Approac for Generating Concurrent Plans in Markov Decision Processes Kasayar Roanimanes kas@cs.umass.edu Sridar Maadevan maadeva@cs.umass.edu Department of Computer Science, University

More information

, 1 1, A complex fraction is a quotient of rational expressions (including their sums) that result

, 1 1, A complex fraction is a quotient of rational expressions (including their sums) that result RT. Complex Fractions Wen working wit algebraic expressions, sometimes we come across needing to simplify expressions like tese: xx 9 xx +, xx + xx + xx, yy xx + xx + +, aa Simplifying Complex Fractions

More information

CE 221 Data Structures and Algorithms

CE 221 Data Structures and Algorithms CE Data Structures and Algoritms Capter 4: Trees (AVL Trees) Text: Read Weiss, 4.4 Izmir University of Economics AVL Trees An AVL (Adelson-Velskii and Landis) tree is a binary searc tree wit a balance

More information

15-122: Principles of Imperative Computation, Summer 2011 Assignment 6: Trees and Secret Codes

15-122: Principles of Imperative Computation, Summer 2011 Assignment 6: Trees and Secret Codes 15-122: Principles of Imperative Computation, Summer 2011 Assignment 6: Trees and Secret Codes William Lovas (wlovas@cs) Karl Naden Out: Tuesday, Friday, June 10, 2011 Due: Monday, June 13, 2011 (Written

More information

Programming Real-Time Applications with Signal

Programming Real-Time Applications with Signal Programming Real-Time Applications wit Signal Paul Le Guernic, Tierry Gautier, Micel Le Borgne, Claude Le Maire To cite tis version: Paul Le Guernic, Tierry Gautier, Micel Le Borgne, Claude Le Maire. Programming

More information

Feature-Based Steganalysis for JPEG Images and its Implications for Future Design of Steganographic Schemes

Feature-Based Steganalysis for JPEG Images and its Implications for Future Design of Steganographic Schemes Feature-Based Steganalysis for JPEG Images and its Implications for Future Design of Steganograpic Scemes Jessica Fridric Dept. of Electrical Engineering, SUNY Bingamton, Bingamton, NY 3902-6000, USA fridric@bingamton.edu

More information

Brief Contributions. A Hybrid Flash File System Based on NOR and NAND Flash Memories for Embedded Devices 1 INTRODUCTION

Brief Contributions. A Hybrid Flash File System Based on NOR and NAND Flash Memories for Embedded Devices 1 INTRODUCTION 1002 IEEE TRANSACTIONS ON COMPUTERS, VOL. 57, NO. 7, JULY 2008 Brief Contributions A Hybrid Flas File System Based on NOR and NAND Flas Memories for Embedded Devices Cul Lee, Student Member, IEEE, Sung

More information

You should be able to visually approximate the slope of a graph. The slope m of the graph of f at the point x, f x is given by

You should be able to visually approximate the slope of a graph. The slope m of the graph of f at the point x, f x is given by Section. Te Tangent Line Problem 89 87. r 5 sin, e, 88. r sin sin Parabola 9 9 Hperbola e 9 9 9 89. 7,,,, 5 7 8 5 ortogonal 9. 5, 5,, 5, 5. Not multiples of eac oter; neiter parallel nor ortogonal 9.,,,

More information

MULTIPLE TOKEN DISTRIBUTED LOOP LOCAL AREA NETWORKS: ANALYSIS

MULTIPLE TOKEN DISTRIBUTED LOOP LOCAL AREA NETWORKS: ANALYSIS ULTIPLE TOKEN DISTRIBUTED LOOP LOCAL AREA NETWORKS: ANALYSIS Nimmagadda Calamaia æ Dept. of CSE JNTU College of Engineering Kakinada, India 533 003. calm@cse.iitkgp.ernet.in Badrinat Ramamurty y Dept.

More information

George Xylomenos and George C. Polyzos. with existing protocols and their eæciency in terms of

George Xylomenos and George C. Polyzos. with existing protocols and their eæciency in terms of IP MULTICASTING FOR WIRELESS MOBILE OSTS George Xylomenos and George C. Polyzos fxgeorge,polyzosg@cs.ucsd.edu Computer Systems Laboratory Department of Computer Science and Engineering University of California,

More information

SlidesGen: Automatic Generation of Presentation Slides for a Technical Paper Using Summarization

SlidesGen: Automatic Generation of Presentation Slides for a Technical Paper Using Summarization Proceedings of te Twenty-Second International FLAIRS Conference (2009) SlidesGen: Automatic Generation of Presentation Slides for a Tecnical Paper Using Summarization M. Sravanti, C. Ravindranat Cowdary

More information

HASH ALGORITHMS: A DESIGN FOR PARALLEL CALCULATIONS

HASH ALGORITHMS: A DESIGN FOR PARALLEL CALCULATIONS HASH ALGORITHMS: A DESIGN FOR PARALLEL CALCULATIONS N.G.Bardis Researc Associate Hellenic Ministry of te Interior, Public Administration and Decentralization 8, Dragatsaniou str., Klatmonos S. 0559, Greece

More information

Limits and Continuity

Limits and Continuity CHAPTER Limits and Continuit. Rates of Cange and Limits. Limits Involving Infinit.3 Continuit.4 Rates of Cange and Tangent Lines An Economic Injur Level (EIL) is a measurement of te fewest number of insect

More information

Network Coding-Aware Queue Management for Unicast Flows over Coded Wireless Networks

Network Coding-Aware Queue Management for Unicast Flows over Coded Wireless Networks Network Coding-Aware Queue Management for Unicast Flows over Coded Wireless Networks Hulya Seferoglu, Atina Markopoulou EECS Dept, University of California, Irvine {seferog, atina}@uci.edu Abstract We

More information

RECONSTRUCTING OF A GIVEN PIXEL S THREE- DIMENSIONAL COORDINATES GIVEN BY A PERSPECTIVE DIGITAL AERIAL PHOTOS BY APPLYING DIGITAL TERRAIN MODEL

RECONSTRUCTING OF A GIVEN PIXEL S THREE- DIMENSIONAL COORDINATES GIVEN BY A PERSPECTIVE DIGITAL AERIAL PHOTOS BY APPLYING DIGITAL TERRAIN MODEL IV. Évfolyam 3. szám - 2009. szeptember Horvát Zoltán orvat.zoltan@zmne.u REONSTRUTING OF GIVEN PIXEL S THREE- DIMENSIONL OORDINTES GIVEN Y PERSPETIVE DIGITL ERIL PHOTOS Y PPLYING DIGITL TERRIN MODEL bsztrakt/bstract

More information

Analytical CHEMISTRY

Analytical CHEMISTRY ISSN : 974-749 Grap kernels and applications in protein classification Jiang Qiangrong*, Xiong Zikang, Zai Can Department of Computer Science, Beijing University of Tecnology, Beijing, (CHINA) E-mail:

More information

On the use of FHT, its modification for practical applications and the structure of Hough image

On the use of FHT, its modification for practical applications and the structure of Hough image On te use of FHT, its modification for practical applications and te structure of Houg image M. Aliev 1,3, E.I. Ersov, D.P. Nikolaev,3 1 Federal Researc Center Computer Science and Control of Russian Academy

More information

Test Generation for Acyclic Sequential Circuits with Hold Registers

Test Generation for Acyclic Sequential Circuits with Hold Registers Test Generation for Acyclic Sequential Circuits wit Hold Registers Tomoo Inoue, Debes Kumar Das, Ciio Sano, Takairo Miara, and Hideo Fujiwara Faculty of Information Sciences Computer Science and Engineering

More information

International Journal of Advanced Research in Computer Science and Software Engineering

International Journal of Advanced Research in Computer Science and Software Engineering Volume 2, Issue 9, September 2012 ISSN: 2277 128X International Journal of Advanced Researc in Computer Science and Software Engineering Researc Paper Available online at: www.ijarcsse.com Performance

More information

Arrays in a Lazy Functional Language a case study: the Fast Fourier Transform

Arrays in a Lazy Functional Language a case study: the Fast Fourier Transform Arrays in a Lazy Functional Language a case study: te Fast Fourier Transform Pieter H. Hartel and Willem G. Vree Department of Computer Systems University of Amsterdam Kruislaan 403, 1098 SJ Amsterdam,

More information

Alternating Direction Implicit Methods for FDTD Using the Dey-Mittra Embedded Boundary Method

Alternating Direction Implicit Methods for FDTD Using the Dey-Mittra Embedded Boundary Method Te Open Plasma Pysics Journal, 2010, 3, 29-35 29 Open Access Alternating Direction Implicit Metods for FDTD Using te Dey-Mittra Embedded Boundary Metod T.M. Austin *, J.R. Cary, D.N. Smite C. Nieter Tec-X

More information

An Analytical Approach to Real-Time Misbehavior Detection in IEEE Based Wireless Networks

An Analytical Approach to Real-Time Misbehavior Detection in IEEE Based Wireless Networks Tis paper was presented as part of te main tecnical program at IEEE INFOCOM 20 An Analytical Approac to Real-Time Misbeavior Detection in IEEE 802. Based Wireless Networks Jin Tang, Yu Ceng Electrical

More information

SEATTLE: A Scalable Ethernet Architecture for Large Enterprises

SEATTLE: A Scalable Ethernet Architecture for Large Enterprises TRC00349 ACM (Typeset by SPi, Manila, Pilippines) 1 of 35 February 23, 2011 14:31 SEATTLE: A Scalable Eternet Arcitecture for Large Enterprises 1 CHANGHOON KIM, Microsoft MATTHEW CAESAR, University of

More information

Cubic smoothing spline

Cubic smoothing spline Cubic smooting spline Menu: QCExpert Regression Cubic spline e module Cubic Spline is used to fit any functional regression curve troug data wit one independent variable x and one dependent random variable

More information

TREES. General Binary Trees The Search Tree ADT Binary Search Trees AVL Trees Threaded trees Splay Trees B-Trees. UNIT -II

TREES. General Binary Trees The Search Tree ADT Binary Search Trees AVL Trees Threaded trees Splay Trees B-Trees. UNIT -II UNIT -II TREES General Binary Trees Te Searc Tree DT Binary Searc Trees VL Trees Treaded trees Splay Trees B-Trees. 2MRKS Q& 1. Define Tree tree is a data structure, wic represents ierarcical relationsip

More information

Efficient Content-Based Indexing of Large Image Databases

Efficient Content-Based Indexing of Large Image Databases Efficient Content-Based Indexing of Large Image Databases ESSAM A. EL-KWAE University of Nort Carolina at Carlotte and MANSUR R. KABUKA University of Miami Large image databases ave emerged in various

More information

M2TECH HIFACE DAC 384KHZ/32BIT DIGITAL-TO-ANALOG CONVERTER USER MANUAL

M2TECH HIFACE DAC 384KHZ/32BIT DIGITAL-TO-ANALOG CONVERTER USER MANUAL M2TECH HIFACE DAC 384KHZ/32BIT DIGITAL-TO-ANALOG CONVERTER USER MANUAL REV. 1.1 5/2013 Warning! Canges or modifications not autorized by te manufacturer can invalidate te compliance to CE regulations and

More information

An Evaluation of Alternative Continuous Media Replication Techniques in Wireless Peer-to-Peer Networks

An Evaluation of Alternative Continuous Media Replication Techniques in Wireless Peer-to-Peer Networks An Evaluation of Alternative Continuous Media eplication Tecniques in Wireless Peer-to-Peer Networks Saram Gandearizade and Tooraj Helmi Department of Computer Science University of Soutern California

More information

Portland State University ECE 588/688. Graphics Processors

Portland State University ECE 588/688. Graphics Processors Portland State University ECE 588/688 Graphics Processors Copyright by Alaa Alameldeen 2018 Why Graphics Processors? Graphics programs have different characteristics from general purpose programs Highly

More information

A Bidirectional Subsethood Based Similarity Measure for Fuzzy Sets

A Bidirectional Subsethood Based Similarity Measure for Fuzzy Sets A Bidirectional Subsetood Based Similarity Measure for Fuzzy Sets Saily Kabir Cristian Wagner Timoty C. Havens and Derek T. Anderson Intelligent Modelling and Analysis (IMA) Group and Lab for Uncertainty

More information

Measuring Length 11and Area

Measuring Length 11and Area Measuring Lengt 11and Area 11.1 Areas of Triangles and Parallelograms 11.2 Areas of Trapezoids, Romuses, and Kites 11.3 Perimeter and Area of Similar Figures 11.4 Circumference and Arc Lengt 11.5 Areas

More information

11. Transceiver Link Debugging Using the System Console

11. Transceiver Link Debugging Using the System Console November 2011 QII53029-11.1.0 11. Transceiver Link Debugging Using te System Console QII53029-11.1.0 Tis capter describes ow to use te Transceiver Toolkit in te Quartus II sotware. Te Transceiver Toolkit

More information

Utilizing Call Admission Control to Derive Optimal Pricing of Multiple Service Classes in Wireless Cellular Networks

Utilizing Call Admission Control to Derive Optimal Pricing of Multiple Service Classes in Wireless Cellular Networks Utilizing Call Admission Control to Derive Optimal Pricing of Multiple Service Classes in Wireless Cellular Networks Okan Yilmaz and Ing-Ray Cen Computer Science Department Virginia Tec {oyilmaz, ircen}@vt.edu

More information

Non-Interferometric Testing

Non-Interferometric Testing NonInterferometric Testing.nb Optics 513 - James C. Wyant 1 Non-Interferometric Testing Introduction In tese notes four non-interferometric tests are described: (1) te Sack-Hartmann test, (2) te Foucault

More information

( ) ( ) Mat 241 Homework Set 5 Due Professor David Schultz. x y. 9 4 The domain is the interior of the hyperbola.

( ) ( ) Mat 241 Homework Set 5 Due Professor David Schultz. x y. 9 4 The domain is the interior of the hyperbola. Mat 4 Homework Set 5 Due Professor David Scultz Directions: Sow all algebraic steps neatly and concisely using proper matematical symbolism. Wen graps and tecnology are to be implemented, do so appropriately.

More information

Search-aware Conditions for Probably Approximately Correct Heuristic Search

Search-aware Conditions for Probably Approximately Correct Heuristic Search Searc-aware Conditions for Probably Approximately Correct Heuristic Searc Roni Stern Ariel Felner Information Systems Engineering Ben Gurion University Beer-Seva, Israel 85104 roni.stern@gmail.com, felner@bgu.ac.il

More information

Introduction to CUDA Algoritmi e Calcolo Parallelo. Daniele Loiacono

Introduction to CUDA Algoritmi e Calcolo Parallelo. Daniele Loiacono Introduction to CUDA Algoritmi e Calcolo Parallelo References This set of slides is mainly based on: CUDA Technical Training, Dr. Antonino Tumeo, Pacific Northwest National Laboratory Slide of Applied

More information

All truths are easy to understand once they are discovered; the point is to discover them. Galileo

All truths are easy to understand once they are discovered; the point is to discover them. Galileo Section 7. olume All truts are easy to understand once tey are discovered; te point is to discover tem. Galileo Te main topic of tis section is volume. You will specifically look at ow to find te volume

More information