COMPUTATION OF DYADIC CONVOLUTION ON GPU FOR EFFICIENT MODELING OF DYADIC LTI SYSTEMS UDC :

Size: px
Start display at page:

Download "COMPUTATION OF DYADIC CONVOLUTION ON GPU FOR EFFICIENT MODELING OF DYADIC LTI SYSTEMS UDC :"

Transcription

1 FACTA UNIVERSITATIS Series: Automatic Cotrol ad Robotics Vol. 10, N o 1, 2011, pp COMPUTATION OF DYADIC CONVOLUTION ON GPU FOR EFFICIENT MODELING OF DYADIC LTI SYSTEMS UDC : GPU Duša B. Gajić, Radomir S. Staković Departmet of Computer Sciece, Faculty of Electroic Egieerig, Uiversity of Niš, Aleksadra Medvedeva 14, Niš, Serbia dule.gajic@gmail.com, radomir.stakovic@gmail.com Abstract. Covolutio systems are a importat class of systems ofte used i may practical applicatios i differet areas of sciece ad egieerig. I the case of systems with biary ecoded iput ad output sigals, fiite dyadic groups are usually assumed as the uderlyig algebraic structure. Modelig ad hadlig of such systems requires frequet computatio of the dyadic covolutio. This paper discusses a method for efficiet computatio of the dyadic covolutio o Graphics Processig Uits (GPUs) with the related algorithm implemeted i Ope Computig Laguage (OpeCL). We cosider several importat issues cocerig the efficiet mappig of the algorithm to the GPU architecture. Performace of the proposed implemetatio is compared with the referet C/C++ implemetatio processed o the Cetral Processig Uit (CPU). Experimetal results cofirm that sigificat speedups are achieved by the applicatio of the proposed GPU calculatio method. Key words: Liear traslatio ivariat (LTI) systems, Dyadic covolutio, Fast Walsh trasform (FWT), GPU parallel programmig, OpeCL 1. INTRODUCTION Liearity, causality, ad traslatio ivariace are features ofte assumed for mathematical models of may systems useful i practice. These features directly lead to covolutio systems, sice ay causal liear traslatio ivariat (LTI) system ca be represeted by a covolutio system. At the same time, ay covolutio system is liear, causal, ad traslatio ivariat. Depedig o the class of sigals to be processed, systems ca be time-ivariat, shift-ivariat, or rotatio-ivariat. For these systems, groups of real umbers ad complex umbers, respectively, are usually assumed as domais for the defiitio of iput ad Received July 06, 2011 Ackowledgemet. The research reported i this paper is partly supported by the Miistry of Educatio ad Sciece of the Republic of Serbia, project OI ( ).

2 60 D. B. GAJIĆ, R. S. STANKOVIĆ output sigals. There are, however, areas where systems o groups differet from these groups are more suitable as mathematical models. I particular, dyadic systems, i.e., systems modeled o dyadic groups (defied bellow), are ofte met i commuicatios, computig, ad cryptography. Dyadic systems were itroduced by Weiser i [36], ad their theory was further cotributed by Harmuth [16, 17], based o the backgroud mathematical theory by Polyak ad Shreider [29]. The dyadic systems have bee itesively studied by Pichler i a series of publicatios [22, 23, 24, 25, 26, 27, 28], ad by several other authors [5, 7]. Recetly, logic etworks, which ca be viewed as facets of digital systems modeled o dyadic groups, were studied from the system theory poit of view [34]. I study ad practical applicatios of dyadic systems, computatio of the dyadic covolutio is frequetly required. Efficiecy of these computatios ofte determies the applicability of such mathematical models i egieerig practice. With that motivatio, i this paper we preset a efficiet techique for a accelerated calculatio of the dyadic covolutio, through a parallel implemetatio of the fast algorithm derived from the covolutio theorem ad processed o the Graphics Processig Uit (GPU). Due to the covolutio theorem, computatio of the dyadic covolutio coverts ito performig two direct ad a iverse Walsh trasform, which ca be doe by the correspodig FFT-like algorithms, i.e., the Fast Walsh Trasform FWT [4, 8, 18]. The proposed implemetatio is developed usig the Ope Computig Laguage (OpeCL) ad processed o a GPU. Experimetal results ad comparisos with the classical implemetatio cofirm that the proposed method leads to a sigificat computatioal speedup. This paper is based o a shorter prelimiary versio preseted at the coferece ICEST 2011 [11], ad it is orgaized as follows. After a itroductio to the ecessary backgroud theory i Sectio 2, Sectio 3 presets a brief review of related work. Sectio 4 is devoted to the descriptio of the mappig of the fast algorithm for computig the dyadic covolutio to the GPU architecture ad the desig of the correspodig OpeCL implemetatio. I Sectio 5, we preset the experimetal eviromet used to evaluate the method, ad report experimetal results. The closig Sectio 6 offers some coclusios draw from the preseted research Dyadic Covolutio 2. BACKGROUND THEORY Covolutio is a mathematical operatio that expresses relatioships betwee values of two sigals (modeled by fuctios f ad g) i poits at a fixed distace. The covolutio C = f gis a fuctio that resembles either the fuctio f or the fuctio g, modified by the other. Whe the fiite dyadic group is used as the algebraic structure o which the covolutio is defied, we use the term dyadic covolutio [4, 18, 30, 32]. The fiite dyadic group of order is defied as C2 C2, where C 2 = ({0,1}, ), ad i1 stads for the additio modulo 2, ad is the direct (Cartesia) product.

3 Computatio of Dyadic Covolutio o GPU for Efficiet Modelig of Dyadic LTI Systems 61 For two fuctios f, g : C2, where is the field of ratioal umbers, the dyadic covolutio is defied as C f g 2 1 ( ) f( x) g( x ), x0 1 0,1,..., 2. (1) I biary otatio, x = (x 1, x 2,, x ), ad = ( 1, 2,, ), where x i, i {0,1}. A direct calculatio of the dyadic covolutio from (1) has the expoetial complexity i the umber of iputs ad is ufeasible i practice for large sigals. Therefore, algorithms for the fast computatio of covolutio are derived by usig the covolutio theorem [18] o the correspodig algebraic structure Walsh Trasform The Walsh trasform [18] is defied by the Walsh matrix W( ) W (1), (2) i1 where deotes the Kroecker product ad 1 1 W (1) 1 1 (3) is the basic Walsh matrix. From (2), W() is geerated by the Kroecker product ad, therefore, the Walsh fuctios (rows of W()) appear i the Hadamard order, see for istace [17]. Sice W() is a self-iverse matrix up to the scalar 2 -, the iverse Walsh trasform is defied as 1 W ( ) 2 W ( ). (4) It follows that both the direct ad the iverse Walsh trasforms ca be computed by usig the same algorithm. The Walsh spectrum S f, h = [S f,h (0), S f,h (1),, S f,h (2-1)] T of a fuctio f : C2, specified by the fuctio vector F = [f (0), f (1),, f (2-1)] T, is defied as S, W( ) F. (5) f h The spectral coefficiets appear i Hadamard orderig, which is idicated by the idex h i S f,h. The fuctio f is recostructed from the Walsh spectrum as F W S (6) 1 2 ( ) f, h, ad (5) ad (6) form the Walsh trasform pair. Example 1. If f : C is specified by the fuctio vector F = [1, 0, 1, 1] T, the 2 2 the correspodig Walsh spectrum S f,h = [S f,h (0), S f,h (1), S f,h (2), S f,h (3)] T ca be calculated directly by usig its defiitio (5)

4 62 D. B. GAJIĆ, R. S. STANKOVIĆ S f, h (1) (2) W. i1 F W F (7) Fast Walsh Trasform The computatio of the Walsh trasform based o its defiitio (5) is iefficiet, sice it expresses the O(N 2 ) time complexity, where N = 2 is the size of the iput vector. Fortuately, a more efficiet algorithm, the Fast Walsh trasform (FWT) with the time complexity of O(NlogN) exist, see for istace [4, 18]. The Fast Walsh trasform ca be defied by usig the followig factorizatio [18] W( ) C w ( ) i i, (8) i1 where j j W(1), i j, Cw ( ) C (1), (1) i w C 1 i w (9) j i I(1), i j. The matrix C ( w ) defies the partial Walsh trasform ad correspods to the i-th step i of the FWT. This particular factorizatio leads to a FFT-like algorithm of the Cooley- Tukey type [12, 18]. 2 Example 2. For the fuctio f : C2 i Example 1, the Walsh spectrum ca be calculated with 4 additios ad 4 subtractios by applyig the FWT algorithm, istead of 16 multiplicatios ad 12 additios i the direct computatio usig the defiitio. This algorithm is derived from the followig factorizatio of the matrix W(2) where S f, h W(2) F ( CC 1 2) F, (10) C1 W(1) I (1), (11) C2 I(1) W (1) (12) Computatios i C 1 ad C 2 ca be performed by the algorithms whose flow-graphs are show i Fig. 1 (a) ad (b). I this figure, the solid ad dashed lies deote additio ad subtractio, respectively.

5 Computatio of Dyadic Covolutio o GPU for Efficiet Modelig of Dyadic LTI Systems 63 (a) Fig. 1 Flow-graphs of the FWT for = 2, (a) First step derived from C 1, (b) Secod step derived from C 2. (b) Therefore, computig the Walsh spectrum goes as follows f (0) S f, h(0) f (1) S f, h(1) F S f, h. (13) f (2) S f, h(2) f (3) S f, h(3) 2.4. Covolutio Theorem I classical Fourier aalysis, the covolutio theorem states that the Fourier trasform of the covolutio fuctio C = f g is the compoetwise product of the Fourier trasforms of f ad g, thus S f g = S f S g, see for istace [4, 8, 18]. I other words, a rather complex operatio i the origial domai coverts ito a simple operatio i the spectral domai. For fuctios o the fiite dyadic group C 2, the computatio of the dyadic covolutio through the applicatio of the covolutio theorem is doe as follows -1 Cf g 2 W ( )(( W( ) F)( W( ) G )). (14) Therefore, a efficiet algorithm for the computatio of the dyadic covolutio ca be developed i terms of the FWT, as illustrated i Fig. 2 for = 2, N = 4. Fig. 2 Computatio of the dyadic covolutio through the applicatio of the covolutio theorem for = 2, N = 4

6 64 D. B. GAJIĆ, R. S. STANKOVIĆ Example 3. For two fuctios f, g : 2 C2, give by their respective fuctio vectors F = [1, 0, 1, 1] T ad G = [0, 1, 0, 1] T, performig the FWT o each of them produces their Walsh spectra S f,h = [3, 1, -1, 1] T ad S g,h = [2, -2, 0, 0] T, respectively. The result of the compoetwise multiplicatio of these two spectra is S f.g,h = [6, -2, 0, 0] T. After the iverse FWT is performed o S f.g,h, followed by multiplicatio with the scalig factor 2-2, we get the dyadic covolutio coefficiets C [1,2,1,2] T. f g 3. BACKGROUND WORK The fast algorithm for the dyadic covolutio is based o the applicatio of the Walsh trasform which is the Fourier trasform o fiite dyadic groups. The implemetatio of various Fast Fourier Trasforms (FFTs) o differet techological platforms is a widely cosidered topic, see for istace [4, 8, 11, 15] ad refereces therei. The calculatio of dyadic covolutio o classical Cetral Processig Uits (CPUs) through the applicatio of the covolutio theorem, both o vectors ad decisio diagrams, is preseted i [30]. Referece [4] presets a applicatio of the dyadic covolutio for the fast multiplicatio of hyper-complex umbers. I recet years, the techique of performig Geeral Purpose computatios o the GPU (GPGPU) has prove to be a suitable approach i solvig may computatioallyitesive tasks [2, 3, 12, 15, 19]. I particular, the GPU-accelerated calculatio of FFT algorithms usig CUDA is described i [15, 20]. Referece [12] presets a OpeCL implemetatio of the FWT that uses the GPU hardware ad leads to sigificat speedups over traditioal CPU processig. However, i our best kowledge, there are o publicatios except the prelimiary versio of this work [11] that discuss either CUDA or OpeCL GPU implemetatios of the fast algorithm for the computatio of dyadic covolutio based o the covolutio theorem. This fact, together with the iteded applicatio of the dyadic covolutio for the modelig of dyadic LTI systems, was the motivatio for the research reported i this paper. 4. MAPPING OF THE ALGORITHM AND IMPLEMENTATION DETAILS The covolutio theorem, expressed i (14), leads to the followig fast algorithm for the calculatio of the dyadic covolutio, illustrated by the example i Fig. 2: Step 1. Perform the FWT o f ad g ad compute their Walsh spectra S f ad S g. Step 2. Perform the compoetwise multiplicatio of S f ad S g. Step 3. Perform the iverse FWT over S f S g to obtai C fg.

7 Computatio of Dyadic Covolutio o GPU for Efficiet Modelig of Dyadic LTI Systems 65 Algorithm 1 FAST CALCULATION OF DYADIC CONVOLUTION C = f * g 1 Allocate buffers buffer1 ad buffer2 i the global memory of the GPU device. 2 Trasfer iput vectors f ad g from the host CPU memory to GPU buffers buffer1 ad buffer2, respectively. 3 Perform the Walsh trasform o vectors stored i buffer1 ad buffer2 usig the followig i-place OpeCL implemetatio of the Cooley-Tukey algorithm for the FWT: a. For each step of the FWT, from step 0 to step (log 2 N) - 1, call the OpeCL kerel for the FWT with iput parameters beig the appropriate buffer i the GPU s global memory ad the value of the curret step 2 step. The kerel is executed by N/2 threads i parallel o the GPU. Each thread reads two elemets, determied by (15) ad (16), from the buffer, performs the operatios defied by the Walsh matrix W (1) ad stores back the results i the same locatios. 4 After computig the FWT of both vectors, execute the OpeCL kerel for the compoetwise multiplicatio of the two Walsh spectra with N threads executed i parallel. The resultig vector is stored i buffer1. 5 Perform the iverse FWT o buffer1 usig the same kerel as for the FWT. 6 Scale the cotets of buffer1 with the factor 2 - usig the OpeCL kerel with N threads executed i parallel. 7 Trasfer the cotets of the GPU buffer buffer1, which holds the resultig dyadic covolutio coefficiets, back to the host CPU memory. Sice both multiplicatio ad compoetwise multiplicatio of vectors ca be doe fast o moder CPUs ad GPUs, the key issue i developig a efficiet implemetatio of the above algorithm is performig the FWT ad the iverse FWT, required i steps 3 ad 5 of Algorithm 1, i a efficiet maer. Therefore, we developed a kerel cotaiig a OpeCL i-place implemetatio of the Cooley-Tukey algorithm for the FWT [12, 18]. As i all FFT-like algorithms, steps of the algorithm are executed sequetially ad parallelism is used oly withi the steps. I each step, N/2 threads are executed i parallel. This large umber of threads permits to overcome the data access latecy to the GPU global memory [2, 3]. Each thread reads two elemets from the GPU buffer with idices op1 ad op2 calculated as op1 thread_id mod step + 2 step (thread_id/step), (15) op2 op1 + step. (16) Parameters thread_id ad step are the global idetifier of the thread ad the idetifier of the curret step of the algorithm, respectively. All threads execute the elemetary butterfly operatio defied by the basic Walsh trasform matrix W(1) ad store the results back i the same locatios i the GPU memory, as i other implemetatios of the iplace FWT algorithms. The compoetwise multiplicatio of vectors is also performed by the correspodig OpeCL kerel which is executed by N threads i parallel, with each thread multiplyig the two correspodig elemets of the iput vectors. After the multiplicatio, the iverse FWT is performed with the same kerel that is used for the direct trasform, followed by scalig with 2 -. The scalig is also performed i parallel through the executio of the correspodig OpeCL kerel.

8 66 D. B. GAJIĆ, R. S. STANKOVIĆ Before executig ay of the kerels, both iput vectors are trasferred from the mai memory to the GPU global memory. After the calculatios, the resultig covolutio coefficiets are copied back to the host. These memory operatios take a sigificat share of the total GPU ruig times as reported i Sectio EXPERIMENTAL ENVIRONMENT AND RESULTS The test platform used to perform the experimets is a HP Pavilio dv7-4060us otebook computer (see Table 1). The OpeCL kerels are developed usig MS Visual Studio 2010 Ultimate ad ATI APP SDK 2.3 [1]. ATI Stream Profiler 2.1 is used for GPU performace aalysis, i accordace with istructios provided i [2]. The source code is compiled for the x64 platform. The referet C/C++ implemetatio is compiled with the maximum level of performace-orieted compiler optimizatios. As i all FFT implemetatios over vectors, the resultig performace is idepedet of the fuctio values. Therefore, we perform the experimets usig radomly geerated biary vectors. All values i Table 2 are a average for 10 program executios. We show the GPU calculatio time ad, also, the time for trasfer of data to/from the GPU which offers a more complete perspective to the reader. To estimate the performace of the proposed method, we performed a comparative aalysis with respect to the classical C/C++ implemetatio of the same algorithm processed o the CPU. Table 1. Specificatio of the experimetal platform. CPU RAM OS GPU - egie speed - global memory - compute uits - processig elemets - price AMD Pheom II N830 triple-core (2.1 GHz) 4GB DDR3 Widows 7 (64-bit) ATI Mobility Radeo MHz 1 GB DDR3 800 MHz ~ 100$ The results of the experimets are preseted i Table 2 ad Fig. 3. The OpeCL implemetatio processed o a iexpesive commodity GPU (Table 1) clearly outperforms the referet CPU implemetatio, by a factor of up to 5.5 whe oly calculatio times are compared, ad by a factor of up to 4.5 whe total times, icludig memory trasfers to/from GPU, are take ito accout. The executio of the same kerels o a more powerful GPU (with more stream cores ad a larger memory badwidth) would directly lead to much larger speedups, which would ot be the case if a more powerful CPU was used for the referet C/C++ implemetatio [3, 12].

9 Computatio of Dyadic Covolutio o GPU for Efficiet Modelig of Dyadic LTI Systems 67 Table 2. Ruig times for the referet CPU ad the proposed GPU implemetatio Time [ms] = log 2 N GPU CPU Computatio Memory time time Number of iput variables CPU Referet C/C++ implemetatio o the CPU GPU Proposed OpeCL C implemetatio o the GPU Fig. 3. Computatio times for the referet CPU ad the proposed GPU implemetatio 6. CONCLUSIONS The dyadic systems are a particular class of liear traslatio ivariat systems o fiite groups modeled i terms of the dyadic covolutio. Efficiet computatio of the dyadic covolutio is therefore essetial i practical applicatios of such systems.

10 68 D. B. GAJIĆ, R. S. STANKOVIĆ The OpeCL parallel algorithm implemetatio for computig the dyadic covolutio, which is processed o the GPU, provides cosiderable speedups, as documeted by experimets over radomly geerated biary fuctios ad a compariso with the referet C/C++ CPU implemetatio of the algorithm. The same OpeCL implemetatio of the dyadic covolutio ca also be used for the fast calculatio of the dyadic autocorrelatio. The proposed method could, therefore, exted the area of applicatios of these operatios to problems where algorithm ruig time is a essetial ad, ofte, limitig factor. REFERENCES 1. AMD Accelerated Parallel Processig SDK, AMD Ic., website last visited o 10/06/ ATI Stream Computig OpeCL Programmig Guide, AMD Ic., T. M. Aamodt, Architectig graphics processors for o-graphics compute acceleratio, i Proc. of the 2009 IEEE Pacific Rim Cof. o Commuicatios, Computers ad Sigal Processig, Special Sessio o Computer Architecture, Victoria, BC, Caada, August 23-26, J. Ardt, Matters Computatioal: Ideas, Algorithms, Source Code, Spriger, S. Coh-Sfetcu, O the theory of liear dyadic ivariat systems, i Proc. Symp. Theory ad Applic. Walsh ad other No-sius. Fuctios, Hatfield, Eglad, S. Coh-Sfetcu, Topics o geeralized covolutio ad Fourier trasform: Theory ad applicatios i digital sigal processig ad system theory, Ph.D. dissertatio, McMaster Uiversity, Hamilto, Otario, Caada, S. Coh-Sfetcu ad S. T. Nichols, O the idetificatio of liear dyadic ivariat systems, IEEE Trasactios o Electromagetic Compatibility, Vol. EMC-17, No. 2, May 1975, J. W. Cooley ad J. W. Tukey, A algorithm for the machie calculatio of complex Fourier series, Mathematics of Computatio, No. 90, 1965, Y. Edow, Walsh harmoizable processes i liear system theory, Cyberetics ad Systems, 1996, Y. Edow ad R. S. Staković, Gibbs derivatives i liear system theory, Cyberetic ad Systems, Vol. 26, 1995, D. B. Gajić ad R. S. Staković, Calculatio of dyadic covolutio usig graphics processig uits ad OpeCL, paper accepted for presetatio at the ICEST 2011 Coferece, Niš, Serbia, Jue 29 - July 1, D. B. Gajić ad R. S. Staković, Computig fast spectral trasforms o graphics processig uits usig OpeCL, i Proc. of the Reed-Muller 2011 Workshop, Tuusula, Filad, May 25-26, 2011, T. H. Frak, Implemetatio of dyadic correlatio, IEEE Tras. Electromagetic Compatibility, Vol. EMC-13, No. 3, 1971, J. E. Gibbs, J. E. Marshall, ad F. R. Pichler, Electrical measuremets i the light of system theory, i IEEE Coferece o the Use of Computers i Measuremet, Uiversity of York, Sept , 1973, ii N. K. Govidaraju et al., High performace discrete Fourier trasforms o graphics processors, i Proc. of the 2008 ACM/IEEE Cof. o Supercomputig, IEEE Press, Austi, Texas, USA, November 15-21, H. F. Harmuth, Die Orthogoalteilug als Verallgemeierug der Zeit-ud Frequezteilug, Archiv der Elektrische Übertragug, vol. 18, 1964, H. F. Harmuth, Sequecy Theory - Foudatios ad Applicatios, Academic Press, New York 1977; Russia editio: MIR, Moscow 1980; reprited i PR Chia - Chiese editio: Publishig House of the People's Post ad Telecommuicatios, Beijig M. G. Karpovsky, R. S. Staković, ad J. T. Astola, Spectral Logic ad Its Applicatios for the Desig of Digital Devices, Wiley-Itersciece, NVidia CUDA: Compute Uified Device Architecture, gpucomputig.html, NVIDIA Corp., website last visited o 21/06/ NVidia CUDA CUFFT Library, NVIDIA Corp., 2007.

11 Computatio of Dyadic Covolutio o GPU for Efficiet Modelig of Dyadic LTI Systems J. Pearl, Optimal dyadic models of time-ivariat systems, IEEE Tras. Comput., Vol. C-24, 1975, F. Pichler, Walsh fuctios ad liear system theory, i Proc. Applic. Walsh Fuctios, Washigto, D.C., 1970, F. Pichler, J. E. Marshall, ad J. E. Gibbs, A system-theory approach to electrical measuremets, i Colloquium o the Theory ad Applicatios of Walsh Fuctios, The Hatfield Polytechic, Hatfield, Hertfordshire, Jue 28-29, F. Pichler, Fast liear methods for image filterig, i Applicatios of Iformatio ad Cotrol Systems, D. G. Laiiotis ad N. S. Tzaues, (Eds.), D. Reidel Publishig Compay, 1980, F. Pichler, Experimets with 1-D ad 2-D sigals usig Gibbs derivatives, i Theory ad Applicatios of Gibbs Derivatives, P. L. Butzer ad R. S. Staković, (Eds.), Belgrade, Serbia: Matematički istitut, 1990, F. Pichler, Some historical remarks o the theory of Walsh fuctios ad their applicatio i iformatio egieerig, i Theory ad Applicatios of Gibbs Derivatives, P. L. Butzer ad R. S. Staković, (Eds.), Belgrade, Serbia: Matematički istitut, F. Pichler, Realizatios of Priggogie's λ-trasform by dyadic covolutio, i Cyberetics ad Systems Research, R. Trappl ad W. Hor, (Eds.), Austria Society for Cyberetic Studies, 1992, ISBN X. 28. F. Pichler, O state-space descriptio of liear dyadic-ivariat systems, i IEEE Trasactios o Electromagetic Compatibility, Vol. EMC-13, No. 3, 1971, B. T. Polyak ad Y. A. Shreider, Applicatios of Walsh fuctios i approximate calculatios, i Voprosy Teorii Matematichskikh Mashi, 1962, M. M. Radmaović, Calculatio of dyadic covolutio through biary decisio diagrams, Facta Uiversitatis: Automatic Cotrol ad Robotics, Vol. 8, No. 1, 2009, R. S. Staković, M. Bhattacharaya, ad J. T. Astola, Calculatio of dyadic autocorrelatio through decisio diagrams, i Proc. of the Europea Coferece o Circuit Theory ad Desig, August 2001, R. S. Staković, M. S. Staković, ad C. Moraga, Remarks o systems ad differetial operators o groups, Facta Uiversitatis: Electroics ad Eergetics, Vol. 18, No. 3, 2005, E. A. Trachteberg ad M. G. Karpovsky, Optimal varyig dyadic structure models of time ivariat systems, i Proc IEEE It. Symp. Circuits ad Systems, Espoo, Filad, Jue 1988, M. A. Thorto, Spectral aalysis of digital logic circuit etlists, i Proc. of the Iteratioal Coferece o Computer Aided Systems Theory (EUROCAST), February 6-11, 2011, The OpeCL Specificatio 1.1, Khroos OpeCL Workig Group, F. E. Weiser, Walsh fuctio aalysis of istataeous oliear stochastic problems, Ph.D. dissertatio, Polytechic Istitute of Brookly, Brookly, Jue IZRAČUNAVANJE DIJADIČKE KONVOLUCIJE NA GRAFIČKIM PROCESNIM JEDINICAMA ZA EFIKASNO MODELOVANJE DIJADIČKIH LTI SISTEMA Duša B. Gajić, Radomir S. Staković Kovolucioi sistemi predstavljaju važu klasu sistema koji se često koriste u mogim praktičim primeama u različitim oblastima auke i ižejerstva. U slučaju kada su ulazi i izlazi sigali biaro kodirai, običo se kao algebarska struktura a kojoj su defiisai odgovarajući sistemi uzimaju koače dijadičke grupe. Modelovaje i rad sa ovakvim sistemima zahtevaju veoma česta izračuavaja dijadičke koovlucije. U ovom radu predstavlje je metod za efikaso izračuavaje dijadičke kovolucije a grafičkim procesim jediicama (Graphics Processig Uits - GPUs) pri čemu je odgovarajući algoritam implemetira primeom programskog jezika Ope Computig Laguage (OpeCL). Razmotreo je ekoliko veoma začajih pitaja koja se tiču efikasog mapiraja algoritma a arhitekturu grafičkih procesih jediica.

12 70 D. B. GAJIĆ, R. S. STANKOVIĆ Performase predložee implemetacije upoređee su sa referetom C/C++ implemetacijom koja se izvršava a cetraloj procesoj jediici (Cetral Processig Uit - CPU). Eksperimetali rezultati potvrđuju da primea predložeog metoda za izračuavaje dijadičke kovolucije a GPU doosi začaja ubrzaja u odosu a referetu CPU implemetaciju. Ključe reči: Lieari traslacioo ezavisi sistemi, Dijadička kovolucija, Brza Walsh-ova trasformacija, GPU paralelo programiraje, OpeCL

Discrete-Time Systems, LTI Systems, and Discrete-Time Convolution

Discrete-Time Systems, LTI Systems, and Discrete-Time Convolution EEL5: Discrete-Time Sigals ad Systems. Itroductio I this set of otes, we begi our mathematical treatmet of discrete-time s. As show i Figure, a discrete-time operates or trasforms some iput sequece x [

More information

Linear time invariant systems

Linear time invariant systems Liear time ivariat systems Alejadro Ribeiro Dept. of Electrical ad Systems Egieerig Uiversity of Pesylvaia aribeiro@seas.upe.edu http://www.seas.upe.edu/users/~aribeiro/ February 25, 2016 Sigal ad Iformatio

More information

Discrete Orthogonal Moment Features Using Chebyshev Polynomials

Discrete Orthogonal Moment Features Using Chebyshev Polynomials Discrete Orthogoal Momet Features Usig Chebyshev Polyomials R. Mukuda, 1 S.H.Og ad P.A. Lee 3 1 Faculty of Iformatio Sciece ad Techology, Multimedia Uiversity 75450 Malacca, Malaysia. Istitute of Mathematical

More information

Orthogonal Gaussian Filters for Signal Processing

Orthogonal Gaussian Filters for Signal Processing Orthogoal Gaussia Filters for Sigal Processig Mark Mackezie ad Kiet Tieu Mechaical Egieerig Uiversity of Wollogog.S.W. Australia Abstract A Gaussia filter usig the Hermite orthoormal series of fuctios

More information

Frequency Domain Filtering

Frequency Domain Filtering Frequecy Domai Filterig Raga Rodrigo October 19, 2010 Outlie Cotets 1 Itroductio 1 2 Fourier Represetatio of Fiite-Duratio Sequeces: The Discrete Fourier Trasform 1 3 The 2-D Discrete Fourier Trasform

More information

Internal Information Representation and Processing

Internal Information Representation and Processing Iteral Iformatio Represetatio ad Processig CSCE 16 - Fudametals of Computer Sciece Dr. Awad Khalil Computer Sciece & Egieerig Departmet The America Uiversity i Cairo Decimal Number System We are used to

More information

Frequency Response of FIR Filters

Frequency Response of FIR Filters EEL335: Discrete-Time Sigals ad Systems. Itroductio I this set of otes, we itroduce the idea of the frequecy respose of LTI systems, ad focus specifically o the frequecy respose of FIR filters.. Steady-state

More information

Chapter 4 : Laplace Transform

Chapter 4 : Laplace Transform 4. Itroductio Laplace trasform is a alterative to solve the differetial equatio by the complex frequecy domai ( s = σ + jω), istead of the usual time domai. The DE ca be easily trasformed ito a algebraic

More information

Optimum LMSE Discrete Transform

Optimum LMSE Discrete Transform Image Trasformatio Two-dimesioal image trasforms are extremely importat areas of study i image processig. The image output i the trasformed space may be aalyzed, iterpreted, ad further processed for implemetig

More information

Chapter 2 Systems and Signals

Chapter 2 Systems and Signals Chapter 2 Systems ad Sigals 1 Itroductio Discrete-Time Sigals: Sequeces Discrete-Time Systems Properties of Liear Time-Ivariat Systems Liear Costat-Coefficiet Differece Equatios Frequecy-Domai Represetatio

More information

Efficient Reverse Converter Design for Five Moduli

Efficient Reverse Converter Design for Five Moduli Joural of Computatios & Modellig, vol., o., 0, 93-08 ISSN: 79-765 (prit), 79-8850 (olie) Iteratioal Scietific ress, 0 Efficiet Reverse Coverter Desig for Five Moduli Set,,,, MohammadReza Taheri, Elham

More information

Discrete-Time Signals and Systems. Signals and Systems. Digital Signals. Discrete-Time Signals. Operations on Sequences: Basic Operations

Discrete-Time Signals and Systems. Signals and Systems. Digital Signals. Discrete-Time Signals. Operations on Sequences: Basic Operations -6.3 Digital Sigal Processig ad Filterig..8 Discrete-ime Sigals ad Systems ime-domai Represetatios of Discrete-ime Sigals ad Systems ime-domai represetatio of a discrete-time sigal as a sequece of umbers

More information

Polynomial Multiplication and Fast Fourier Transform

Polynomial Multiplication and Fast Fourier Transform Polyomial Multiplicatio ad Fast Fourier Trasform Com S 477/577 Notes Ya-Bi Jia Sep 19, 2017 I this lecture we will describe the famous algorithm of fast Fourier trasform FFT, which has revolutioized digital

More information

ECE-S352 Introduction to Digital Signal Processing Lecture 3A Direct Solution of Difference Equations

ECE-S352 Introduction to Digital Signal Processing Lecture 3A Direct Solution of Difference Equations ECE-S352 Itroductio to Digital Sigal Processig Lecture 3A Direct Solutio of Differece Equatios Discrete Time Systems Described by Differece Equatios Uit impulse (sample) respose h() of a DT system allows

More information

2D DSP Basics: 2D Systems

2D DSP Basics: 2D Systems - Digital Image Processig ad Compressio D DSP Basics: D Systems D Systems T[ ] y = T [ ] Liearity Additivity: If T y = T [ ] The + T y = y + y Homogeeity: If The T y = T [ ] a T y = ay = at [ ] Liearity

More information

The target reliability and design working life

The target reliability and design working life Safety ad Security Egieerig IV 161 The target reliability ad desig workig life M. Holický Kloker Istitute, CTU i Prague, Czech Republic Abstract Desig workig life ad target reliability levels recommeded

More information

Applications of Distributed Arithmetic to Digital Signal Processing: A Tutorial Review

Applications of Distributed Arithmetic to Digital Signal Processing: A Tutorial Review pplicatios of Distriuted rithmetic to Digital Sigal Processig: Tutorial Review Ref: Staley. White, pplicatios of Distriuted rithmetic to Digital Sigal Processig: Tutorial Review, IEEE SSP Magazie, July,

More information

The z-transform. 7.1 Introduction. 7.2 The z-transform Derivation of the z-transform: x[n] = z n LTI system, h[n] z = re j

The z-transform. 7.1 Introduction. 7.2 The z-transform Derivation of the z-transform: x[n] = z n LTI system, h[n] z = re j The -Trasform 7. Itroductio Geeralie the complex siusoidal represetatio offered by DTFT to a represetatio of complex expoetial sigals. Obtai more geeral characteristics for discrete-time LTI systems. 7.

More information

Classification of problem & problem solving strategies. classification of time complexities (linear, logarithmic etc)

Classification of problem & problem solving strategies. classification of time complexities (linear, logarithmic etc) Classificatio of problem & problem solvig strategies classificatio of time complexities (liear, arithmic etc) Problem subdivisio Divide ad Coquer strategy. Asymptotic otatios, lower boud ad upper boud:

More information

Spectral Observer with Reduced Information Demand

Spectral Observer with Reduced Information Demand Spectral Observer with Reduced Iformatio Demad Gy. Orosz, L. Sujbert, G. Péceli Departmet of Measuremet ad Iformatio Systems, Budapest Uiversity of Techology ad Ecoomics Magyar tudósok krt. 2., -52 Budapest,

More information

Parallel Vector Algorithms David A. Padua

Parallel Vector Algorithms David A. Padua Parallel Vector Algorithms 1 of 32 Itroductio Next, we study several algorithms where parallelism ca be easily expressed i terms of array operatios. We will use Fortra 90 to represet these algorithms.

More information

Olli Simula T / Chapter 1 3. Olli Simula T / Chapter 1 5

Olli Simula T / Chapter 1 3. Olli Simula T / Chapter 1 5 Sigals ad Systems Sigals ad Systems Sigals are variables that carry iformatio Systemstake sigals as iputs ad produce sigals as outputs The course deals with the passage of sigals through systems T-6.4

More information

EE / EEE SAMPLE STUDY MATERIAL. GATE, IES & PSUs Signal System. Electrical Engineering. Postal Correspondence Course

EE / EEE SAMPLE STUDY MATERIAL. GATE, IES & PSUs Signal System. Electrical Engineering. Postal Correspondence Course Sigal-EE Postal Correspodece Course 1 SAMPLE STUDY MATERIAL Electrical Egieerig EE / EEE Postal Correspodece Course GATE, IES & PSUs Sigal System Sigal-EE Postal Correspodece Course CONTENTS 1. SIGNAL

More information

Signals & Systems Chapter3

Signals & Systems Chapter3 Sigals & Systems Chapter3 1.2 Discrete-Time (D-T) Sigals Electroic systems do most of the processig of a sigal usig a computer. A computer ca t directly process a C-T sigal but istead eeds a stream of

More information

A New Class of Ternary Zero Correlation Zone Sequence Sets Based on Mutually Orthogonal Complementary Sets

A New Class of Ternary Zero Correlation Zone Sequence Sets Based on Mutually Orthogonal Complementary Sets IOSR Joural of Electroics ad Commuicatio Egieerig (IOSR-JECE) e-issn: 2278-2834,p- ISSN: 2278-8735.Volume 0, Issue 3, Ver. I (May - Ju.205), PP 08-3 www.iosrjourals.org A New Class of Terary Zero Correlatio

More information

Digital Signal Processing, Fall 2006

Digital Signal Processing, Fall 2006 Digital Sigal Processig, Fall 26 Lecture 1: Itroductio, Discrete-time sigals ad systems Zheg-Hua Ta Departmet of Electroic Systems Aalborg Uiversity, Demark zt@kom.aau.dk 1 Part I: Itroductio Itroductio

More information

Sensitivity Analysis of Daubechies 4 Wavelet Coefficients for Reduction of Reconstructed Image Error

Sensitivity Analysis of Daubechies 4 Wavelet Coefficients for Reduction of Reconstructed Image Error Proceedigs of the 6th WSEAS Iteratioal Coferece o SIGNAL PROCESSING, Dallas, Texas, USA, March -4, 7 67 Sesitivity Aalysis of Daubechies 4 Wavelet Coefficiets for Reductio of Recostructed Image Error DEVINDER

More information

Module 18 Discrete Time Signals and Z-Transforms Objective: Introduction : Description: Discrete Time Signal representation

Module 18 Discrete Time Signals and Z-Transforms Objective: Introduction : Description: Discrete Time Signal representation Module 8 Discrete Time Sigals ad Z-Trasforms Objective:To uderstad represetig discrete time sigals, apply z trasform for aalyzigdiscrete time sigals ad to uderstad the relatio to Fourier trasform Itroductio

More information

Image Segmentation on Spiral Architecture

Image Segmentation on Spiral Architecture Image Segmetatio o Spiral Architecture Qiag Wu, Xiagjia He ad Tom Hitz Departmet of Computer Systems Uiversity of Techology, Sydey PO Box 3, Broadway Street, Sydey 007, Australia {wuq,sea,hitz}@it.uts.edu.au

More information

ADVANCED DIGITAL SIGNAL PROCESSING

ADVANCED DIGITAL SIGNAL PROCESSING ADVANCED DIGITAL SIGNAL PROCESSING PROF. S. C. CHAN (email : sccha@eee.hku.hk, Rm. CYC-702) DISCRETE-TIME SIGNALS AND SYSTEMS MULTI-DIMENSIONAL SIGNALS AND SYSTEMS RANDOM PROCESSES AND APPLICATIONS ADAPTIVE

More information

IP Reference guide for integer programming formulations.

IP Reference guide for integer programming formulations. IP Referece guide for iteger programmig formulatios. by James B. Orli for 15.053 ad 15.058 This documet is iteded as a compact (or relatively compact) guide to the formulatio of iteger programs. For more

More information

Interval Intuitionistic Trapezoidal Fuzzy Prioritized Aggregating Operators and their Application to Multiple Attribute Decision Making

Interval Intuitionistic Trapezoidal Fuzzy Prioritized Aggregating Operators and their Application to Multiple Attribute Decision Making Iterval Ituitioistic Trapezoidal Fuzzy Prioritized Aggregatig Operators ad their Applicatio to Multiple Attribute Decisio Makig Xia-Pig Jiag Chogqig Uiversity of Arts ad Scieces Chia cqmaagemet@163.com

More information

Solution of Differential Equation from the Transform Technique

Solution of Differential Equation from the Transform Technique Iteratioal Joural of Computatioal Sciece ad Mathematics ISSN 0974-3189 Volume 3, Number 1 (2011), pp 121-125 Iteratioal Research Publicatio House http://wwwirphousecom Solutio of Differetial Equatio from

More information

Proceedings of the 2nd International Conference on Computer Science and Electronics Engineering (ICCSEE 2013)

Proceedings of the 2nd International Conference on Computer Science and Electronics Engineering (ICCSEE 2013) Proceedigs of the 2d Iteratioal Coferece o Computer Sciece ad Electroics Egieerig (ICCE 20) GPU Implemetatio of the Electric Power System Model for Real-Time Simulatio of Electromagetic Trasiets Slawomir

More information

A Block Cipher Using Linear Congruences

A Block Cipher Using Linear Congruences Joural of Computer Sciece 3 (7): 556-560, 2007 ISSN 1549-3636 2007 Sciece Publicatios A Block Cipher Usig Liear Cogrueces 1 V.U.K. Sastry ad 2 V. Jaaki 1 Academic Affairs, Sreeidhi Istitute of Sciece &

More information

Time-Domain Representations of LTI Systems

Time-Domain Representations of LTI Systems 2.1 Itroductio Objectives: 1. Impulse resposes of LTI systems 2. Liear costat-coefficiets differetial or differece equatios of LTI systems 3. Bloc diagram represetatios of LTI systems 4. State-variable

More information

ON THE EXISTENCE OF A GROUP ORTHONORMAL BASIS. Peter Zizler (Received 20 June, 2014)

ON THE EXISTENCE OF A GROUP ORTHONORMAL BASIS. Peter Zizler (Received 20 June, 2014) NEW ZEALAND JOURNAL OF MATHEMATICS Volume 45 (2015, 45-52 ON THE EXISTENCE OF A GROUP ORTHONORMAL BASIS Peter Zizler (Received 20 Jue, 2014 Abstract. Let G be a fiite group ad let l 2 (G be a fiite dimesioal

More information

Discrete-Time Signals and Systems. Discrete-Time Signals and Systems. Signal Symmetry. Elementary Discrete-Time Signals.

Discrete-Time Signals and Systems. Discrete-Time Signals and Systems. Signal Symmetry. Elementary Discrete-Time Signals. Discrete-ime Sigals ad Systems Discrete-ime Sigals ad Systems Dr. Deepa Kudur Uiversity of oroto Referece: Sectios. -.5 of Joh G. Proakis ad Dimitris G. Maolakis, Digital Sigal Processig: Priciples, Algorithms,

More information

A Unified Approach on Fast Training of Feedforward and Recurrent Networks Using EM Algorithm

A Unified Approach on Fast Training of Feedforward and Recurrent Networks Using EM Algorithm 2270 IEEE TRASACTIOS O SIGAL PROCESSIG, VOL. 46, O. 8, AUGUST 1998 [12] Q. T. Zhag, K. M. Wog, P. C. Yip, ad J. P. Reilly, Statistical aalysis of the performace of iformatio criteria i the detectio of

More information

Signal Processing in Mechatronics

Signal Processing in Mechatronics Sigal Processig i Mechatroics Zhu K.P. AIS, UM. Lecture, Brief itroductio to Sigals ad Systems, Review of Liear Algebra ad Sigal Processig Related Mathematics . Brief Itroductio to Sigals What is sigal

More information

Oblivious Transfer using Elliptic Curves

Oblivious Transfer using Elliptic Curves Oblivious Trasfer usig Elliptic Curves bhishek Parakh Louisiaa State Uiversity, ato Rouge, L May 4, 006 bstract: This paper proposes a algorithm for oblivious trasfer usig elliptic curves lso, we preset

More information

Chandrasekhar Type Algorithms. for the Riccati Equation of Lainiotis Filter

Chandrasekhar Type Algorithms. for the Riccati Equation of Lainiotis Filter Cotemporary Egieerig Scieces, Vol. 3, 00, o. 4, 9-00 Chadrasekhar ype Algorithms for the Riccati Equatio of Laiiotis Filter Nicholas Assimakis Departmet of Electroics echological Educatioal Istitute of

More information

Let A(x) and B(x) be two polynomials of degree n 1:

Let A(x) and B(x) be two polynomials of degree n 1: MI-EVY (2011/2012) J. Holub: 4. DFT, FFT ad Patter Matchig p. 2/42 Operatios o polyomials MI-EVY (2011/2012) J. Holub: 4. DFT, FFT ad Patter Matchig p. 4/42 Efficiet Patter Matchig (MI-EVY) 4. DFT, FFT

More information

DIGITAL MEASUREMENT OF POWER SYSTEM HARMONIC MAGNITUDE AND PHASE ANGLE

DIGITAL MEASUREMENT OF POWER SYSTEM HARMONIC MAGNITUDE AND PHASE ANGLE DIGIL MESUREMEN OF POWER SYSEM HRMONIC MGNIUDE ND PHSE NGLE R Micheletti (, R Pieri ( ( Departmet of Electrical Systems ad utomatio, Uiversity of Pisa, Via Diotisalvi, I-566 Pisa, Italy Phoe +39 5 565,

More information

The Discrete Fourier Transform

The Discrete Fourier Transform The iscrete Fourier Trasform The discrete-time Fourier trasform (TFT) of a sequece is a cotiuous fuctio of!, ad repeats with period. I practice we usually wat to obtai the Fourier compoets usig digital

More information

Warped, Chirp Z-Transform: Radar Signal Processing

Warped, Chirp Z-Transform: Radar Signal Processing arped, Chirp Z-Trasform: Radar Sigal Processig by Garimella Ramamurthy Report o: IIIT/TR// Cetre for Commuicatios Iteratioal Istitute of Iformatio Techology Hyderabad - 5 3, IDIA Jauary ARPED, CHIRP Z

More information

Enterprise interoperability measurement - Basic concepts

Enterprise interoperability measurement - Basic concepts Eterprise iteroperability measuremet - Basic cocepts Nicolas Dacli, David Che, Bruo Vallespir LAPS/GRAI, Uiversity Bordeaux 1, ENSEIRB, UMR 5131 CNRS, 351 cours de la Libératio, 33405 Talece Cedex, Frace

More information

Module 5 EMBEDDED WAVELET CODING. Version 2 ECE IIT, Kharagpur

Module 5 EMBEDDED WAVELET CODING. Version 2 ECE IIT, Kharagpur Module 5 EMBEDDED WAVELET CODING Versio ECE IIT, Kharagpur Lesso 4 SPIHT algorithm Versio ECE IIT, Kharagpur Istructioal Objectives At the ed of this lesso, the studets should be able to:. State the limitatios

More information

A collocation method for singular integral equations with cosecant kernel via Semi-trigonometric interpolation

A collocation method for singular integral equations with cosecant kernel via Semi-trigonometric interpolation Iteratioal Joural of Mathematics Research. ISSN 0976-5840 Volume 9 Number 1 (017) pp. 45-51 Iteratioal Research Publicatio House http://www.irphouse.com A collocatio method for sigular itegral equatios

More information

Chapter 7: The z-transform. Chih-Wei Liu

Chapter 7: The z-transform. Chih-Wei Liu Chapter 7: The -Trasform Chih-Wei Liu Outlie Itroductio The -Trasform Properties of the Regio of Covergece Properties of the -Trasform Iversio of the -Trasform The Trasfer Fuctio Causality ad Stability

More information

Comparison Study of Series Approximation. and Convergence between Chebyshev. and Legendre Series

Comparison Study of Series Approximation. and Convergence between Chebyshev. and Legendre Series Applied Mathematical Scieces, Vol. 7, 03, o. 6, 3-337 HIKARI Ltd, www.m-hikari.com http://d.doi.org/0.988/ams.03.3430 Compariso Study of Series Approimatio ad Covergece betwee Chebyshev ad Legedre Series

More information

Finite-length Discrete Transforms. Chapter 5, Sections

Finite-length Discrete Transforms. Chapter 5, Sections Fiite-legth Discrete Trasforms Chapter 5, Sectios 5.2-50 5.0 Dr. Iyad djafar Outlie The Discrete Fourier Trasform (DFT) Matrix Represetatio of DFT Fiite-legth Sequeces Circular Covolutio DFT Symmetry Properties

More information

ELEC1200: A System View of Communications: from Signals to Packets Lecture 3

ELEC1200: A System View of Communications: from Signals to Packets Lecture 3 ELEC2: A System View of Commuicatios: from Sigals to Packets Lecture 3 Commuicatio chaels Discrete time Chael Modelig the chael Liear Time Ivariat Systems Step Respose Respose to sigle bit Respose to geeral

More information

Seed and Sieve of Odd Composite Numbers with Applications in Factorization of Integers

Seed and Sieve of Odd Composite Numbers with Applications in Factorization of Integers IOSR Joural of Mathematics (IOSR-JM) e-issn: 78-578, p-issn: 319-75X. Volume 1, Issue 5 Ver. VIII (Sep. - Oct.01), PP 01-07 www.iosrjourals.org Seed ad Sieve of Odd Composite Numbers with Applicatios i

More information

Exam. Notes: A single A4 sheet of paper (double sided; hand-written or computer typed)

Exam. Notes: A single A4 sheet of paper (double sided; hand-written or computer typed) Exam February 8th, 8 Sigals & Systems (5-575-) Prof. R. D Adrea Exam Exam Duratio: 5 Mi Number of Problems: 5 Number of Poits: 5 Permitted aids: Importat: Notes: A sigle A sheet of paper (double sided;

More information

Lainiotis filter implementation. via Chandrasekhar type algorithm

Lainiotis filter implementation. via Chandrasekhar type algorithm Joural of Computatios & Modellig, vol.1, o.1, 2011, 115-130 ISSN: 1792-7625 prit, 1792-8850 olie Iteratioal Scietific Press, 2011 Laiiotis filter implemetatio via Chadrasehar type algorithm Nicholas Assimais

More information

Fast Gauss transforms with complex parameters using NFFTs

Fast Gauss transforms with complex parameters using NFFTs Fast Gauss trasforms with complex parameters usig NFFTs Stefa Kuis Daiel Potts Gabriele Steidl Jue 3, 5 We costruct a fast algorithm for the computatio of discrete Gauss trasforms with complex parameters,

More information

Basics of Probability Theory (for Theory of Computation courses)

Basics of Probability Theory (for Theory of Computation courses) Basics of Probability Theory (for Theory of Computatio courses) Oded Goldreich Departmet of Computer Sciece Weizma Istitute of Sciece Rehovot, Israel. oded.goldreich@weizma.ac.il November 24, 2008 Preface.

More information

PAijpam.eu ON TENSOR PRODUCT DECOMPOSITION

PAijpam.eu ON TENSOR PRODUCT DECOMPOSITION Iteratioal Joural of Pure ad Applied Mathematics Volume 103 No 3 2015, 537-545 ISSN: 1311-8080 (prited versio); ISSN: 1314-3395 (o-lie versio) url: http://wwwijpameu doi: http://dxdoiorg/1012732/ijpamv103i314

More information

Practical Spectral Anaysis (continue) (from Boaz Porat s book) Frequency Measurement

Practical Spectral Anaysis (continue) (from Boaz Porat s book) Frequency Measurement Practical Spectral Aaysis (cotiue) (from Boaz Porat s book) Frequecy Measuremet Oe of the most importat applicatios of the DFT is the measuremet of frequecies of periodic sigals (eg., siusoidal sigals),

More information

AN OPEN-PLUS-CLOSED-LOOP APPROACH TO SYNCHRONIZATION OF CHAOTIC AND HYPERCHAOTIC MAPS

AN OPEN-PLUS-CLOSED-LOOP APPROACH TO SYNCHRONIZATION OF CHAOTIC AND HYPERCHAOTIC MAPS http://www.paper.edu.c Iteratioal Joural of Bifurcatio ad Chaos, Vol. 1, No. 5 () 119 15 c World Scietific Publishig Compay AN OPEN-PLUS-CLOSED-LOOP APPROACH TO SYNCHRONIZATION OF CHAOTIC AND HYPERCHAOTIC

More information

Question1 Multiple choices (circle the most appropriate one):

Question1 Multiple choices (circle the most appropriate one): Philadelphia Uiversity Studet Name: Faculty of Egieerig Studet Number: Dept. of Computer Egieerig Fial Exam, First Semester: 2014/2015 Course Title: Digital Sigal Aalysis ad Processig Date: 01/02/2015

More information

Modified Logistic Maps for Cryptographic Application

Modified Logistic Maps for Cryptographic Application Applied Mathematics, 25, 6, 773-782 Published Olie May 25 i SciRes. http://www.scirp.org/joural/am http://dx.doi.org/.4236/am.25.6573 Modified Logistic Maps for Cryptographic Applicatio Shahram Etemadi

More information

SCALING OF NUMBERS IN RESIDUE ARITHMETIC WITH THE FLEXIBLE SELECTION OF SCALING FACTOR

SCALING OF NUMBERS IN RESIDUE ARITHMETIC WITH THE FLEXIBLE SELECTION OF SCALING FACTOR POZNAN UNIVE RSITY OF TE CHNOLOGY ACADE MIC JOURNALS No 76 Electrical Egieerig 203 Zeo ULMAN* Macie CZYŻAK* Robert SMYK* SCALING OF NUMBERS IN RESIDUE ARITHMETIC WITH THE FLEXIBLE SELECTION OF SCALING

More information

Research Article A Unified Weight Formula for Calculating the Sample Variance from Weighted Successive Differences

Research Article A Unified Weight Formula for Calculating the Sample Variance from Weighted Successive Differences Discrete Dyamics i Nature ad Society Article ID 210761 4 pages http://dxdoiorg/101155/2014/210761 Research Article A Uified Weight Formula for Calculatig the Sample Variace from Weighted Successive Differeces

More information

ADVANCED SOFTWARE ENGINEERING

ADVANCED SOFTWARE ENGINEERING ADVANCED SOFTWARE ENGINEERING COMP 3705 Exercise Usage-based Testig ad Reliability Versio 1.0-040406 Departmet of Computer Ssciece Sada Narayaappa, Aeliese Adrews Versio 1.1-050405 Departmet of Commuicatio

More information

Complex Algorithms for Lattice Adaptive IIR Notch Filter

Complex Algorithms for Lattice Adaptive IIR Notch Filter 4th Iteratioal Coferece o Sigal Processig Systems (ICSPS ) IPCSIT vol. 58 () () IACSIT Press, Sigapore DOI:.7763/IPCSIT..V58. Complex Algorithms for Lattice Adaptive IIR Notch Filter Hog Liag +, Nig Jia

More information

ELEG3503 Introduction to Digital Signal Processing

ELEG3503 Introduction to Digital Signal Processing ELEG3503 Itroductio to Digital Sigal Processig 1 Itroductio 2 Basics of Sigals ad Systems 3 Fourier aalysis 4 Samplig 5 Liear time-ivariat (LTI) systems 6 z-trasform 7 System Aalysis 8 System Realizatio

More information

Pipelined and Parallel Recursive and Adaptive Filters

Pipelined and Parallel Recursive and Adaptive Filters VLSI Digital Sigal Processig Systems Pipelied ad Parallel Recursive ad Adaptive Filters La-Da Va 范倫達, Ph. D. Departmet of Computer Sciece Natioal Chiao ug Uiversity aiwa, R.O.C. Fall, 05 ldva@cs.ctu.edu.tw

More information

Similarity Solutions to Unsteady Pseudoplastic. Flow Near a Moving Wall

Similarity Solutions to Unsteady Pseudoplastic. Flow Near a Moving Wall Iteratioal Mathematical Forum, Vol. 9, 04, o. 3, 465-475 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/0.988/imf.04.48 Similarity Solutios to Usteady Pseudoplastic Flow Near a Movig Wall W. Robi Egieerig

More information

Fall 2011, EE123 Digital Signal Processing

Fall 2011, EE123 Digital Signal Processing Lecture 5 Miki Lustig, UCB September 14, 211 Miki Lustig, UCB Motivatios for Discrete Fourier Trasform Sampled represetatio i time ad frequecy umerical Fourier aalysis requires a Fourier represetatio that

More information

CSI 5163 (95.573) ALGORITHM ANALYSIS AND DESIGN

CSI 5163 (95.573) ALGORITHM ANALYSIS AND DESIGN CSI 5163 (95.573) ALGORITHM ANALYSIS AND DESIGN CSI 5163 (95.5703) ALGORITHM ANALYSIS AND DESIGN (3 cr.) (T) Topics of curret iterest i the desig ad aalysis of computer algorithms for graphtheoretical

More information

Some Explicit Formulae of NAF and its Left-to-Right. Analogue Based on Booth Encoding

Some Explicit Formulae of NAF and its Left-to-Right. Analogue Based on Booth Encoding Vol.7, No.6 (01, pp.69-74 http://dx.doi.org/10.1457/ijsia.01.7.6.7 Some Explicit Formulae of NAF ad its Left-to-Right Aalogue Based o Booth Ecodig Dog-Guk Ha, Okyeo Yi, ad Tsuyoshi Takagi Kookmi Uiversity,

More information

subject to A 1 x + A 2 y b x j 0, j = 1,,n 1 y j = 0 or 1, j = 1,,n 2

subject to A 1 x + A 2 y b x j 0, j = 1,,n 1 y j = 0 or 1, j = 1,,n 2 Additioal Brach ad Boud Algorithms 0-1 Mixed-Iteger Liear Programmig The brach ad boud algorithm described i the previous sectios ca be used to solve virtually all optimizatio problems cotaiig iteger variables,

More information

A new iterative algorithm for reconstructing a signal from its dyadic wavelet transform modulus maxima

A new iterative algorithm for reconstructing a signal from its dyadic wavelet transform modulus maxima ol 46 No 6 SCIENCE IN CHINA (Series F) December 3 A ew iterative algorithm for recostructig a sigal from its dyadic wavelet trasform modulus maxima ZHANG Zhuosheg ( u ), LIU Guizhog ( q) & LIU Feg ( )

More information

G r a d e 1 1 P r e - C a l c u l u s M a t h e m a t i c s ( 3 0 S )

G r a d e 1 1 P r e - C a l c u l u s M a t h e m a t i c s ( 3 0 S ) G r a d e 1 1 P r e - C a l c u l u s M a t h e m a t i c s ( 3 0 S ) Grade 11 Pre-Calculus Mathematics (30S) is desiged for studets who ited to study calculus ad related mathematics as part of post-secodary

More information

6.003 Homework #3 Solutions

6.003 Homework #3 Solutions 6.00 Homework # Solutios Problems. Complex umbers a. Evaluate the real ad imagiary parts of j j. π/ Real part = Imagiary part = 0 e Euler s formula says that j = e jπ/, so jπ/ j π/ j j = e = e. Thus the

More information

Chapter 7 z-transform

Chapter 7 z-transform Chapter 7 -Trasform Itroductio Trasform Uilateral Trasform Properties Uilateral Trasform Iversio of Uilateral Trasform Determiig the Frequecy Respose from Poles ad Zeros Itroductio Role i Discrete-Time

More information

FIR Filter Design: Part II

FIR Filter Design: Part II EEL335: Discrete-Time Sigals ad Systems. Itroductio I this set of otes, we cosider how we might go about desigig FIR filters with arbitrary frequecy resposes, through compositio of multiple sigle-peak

More information

Filter banks. Separately, the lowpass and highpass filters are not invertible. removes the highest frequency 1/ 2and

Filter banks. Separately, the lowpass and highpass filters are not invertible. removes the highest frequency 1/ 2and Filter bas Separately, the lowpass ad highpass filters are ot ivertible T removes the highest frequecy / ad removes the lowest frequecy Together these filters separate the sigal ito low-frequecy ad high-frequecy

More information

Robust Resource Allocation in Parallel and Distributed Computing Systems (tentative)

Robust Resource Allocation in Parallel and Distributed Computing Systems (tentative) Robust Resource Allocatio i Parallel ad Distributed Computig Systems (tetative) Ph.D. cadidate V. Shestak Colorado State Uiversity Electrical ad Computer Egieerig Departmet Fort Collis, Colorado, USA shestak@colostate.edu

More information

Chimica Inorganica 3

Chimica Inorganica 3 himica Iorgaica Irreducible Represetatios ad haracter Tables Rather tha usig geometrical operatios, it is ofte much more coveiet to employ a ew set of group elemets which are matrices ad to make the rule

More information

Lecture 3: Divide and Conquer: Fast Fourier Transform

Lecture 3: Divide and Conquer: Fast Fourier Transform Lecture 3: Divide ad Coquer: Fast Fourier Trasform Polyomial Operatios vs. Represetatios Divide ad Coquer Algorithm Collapsig Samples / Roots of Uity FFT, IFFT, ad Polyomial Multiplicatio Polyomial operatios

More information

EE260: Digital Design, Spring n MUX Gate n Rudimentary functions n Binary Decoders. n Binary Encoders n Priority Encoders

EE260: Digital Design, Spring n MUX Gate n Rudimentary functions n Binary Decoders. n Binary Encoders n Priority Encoders EE260: Digital Desig, Sprig 2018 EE 260: Itroductio to Digital Desig MUXs, Ecoders, Decoders Yao Zheg Departmet of Electrical Egieerig Uiversity of Hawaiʻi at Māoa Overview of Ecoder ad Decoder MUX Gate

More information

Live Line Measuring the Parameters of 220 kv Transmission Lines with Mutual Inductance in Hainan Power Grid

Live Line Measuring the Parameters of 220 kv Transmission Lines with Mutual Inductance in Hainan Power Grid Egieerig, 213, 5, 146-151 doi:1.4236/eg.213.51b27 Published Olie Jauary 213 (http://www.scirp.org/joural/eg) Live Lie Measurig the Parameters of 22 kv Trasmissio Lies with Mutual Iductace i Haia Power

More information

Principle Of Superposition

Principle Of Superposition ecture 5: PREIMINRY CONCEP O RUCUR NYI Priciple Of uperpositio Mathematically, the priciple of superpositio is stated as ( a ) G( a ) G( ) G a a or for a liear structural system, the respose at a give

More information

An Intuitionistic fuzzy count and cardinality of Intuitionistic fuzzy sets

An Intuitionistic fuzzy count and cardinality of Intuitionistic fuzzy sets Malaya Joural of Matematik 4(1)(2013) 123 133 A Ituitioistic fuzzy cout ad cardiality of Ituitioistic fuzzy sets B. K. Tripathy a, S. P. Jea b ad S. K. Ghosh c, a School of Computig Scieces ad Egieerig,

More information

DECOMPOSITION METHOD FOR SOLVING A SYSTEM OF THIRD-ORDER BOUNDARY VALUE PROBLEMS. Park Road, Islamabad, Pakistan

DECOMPOSITION METHOD FOR SOLVING A SYSTEM OF THIRD-ORDER BOUNDARY VALUE PROBLEMS. Park Road, Islamabad, Pakistan Mathematical ad Computatioal Applicatios, Vol. 9, No. 3, pp. 30-40, 04 DECOMPOSITION METHOD FOR SOLVING A SYSTEM OF THIRD-ORDER BOUNDARY VALUE PROBLEMS Muhammad Aslam Noor, Khalida Iayat Noor ad Asif Waheed

More information

Signal Processing. Lecture 02: Discrete Time Signals and Systems. Ahmet Taha Koru, Ph. D. Yildiz Technical University.

Signal Processing. Lecture 02: Discrete Time Signals and Systems. Ahmet Taha Koru, Ph. D. Yildiz Technical University. Sigal Processig Lecture 02: Discrete Time Sigals ad Systems Ahmet Taha Koru, Ph. D. Yildiz Techical Uiversity 2017-2018 Fall ATK (YTU) Sigal Processig 2017-2018 Fall 1 / 51 Discrete Time Sigals Discrete

More information

Scheduling under Uncertainty using MILP Sensitivity Analysis

Scheduling under Uncertainty using MILP Sensitivity Analysis Schedulig uder Ucertaity usig MILP Sesitivity Aalysis M. Ierapetritou ad Zheya Jia Departmet of Chemical & Biochemical Egieerig Rutgers, the State Uiversity of New Jersey Piscataway, NJ Abstract The aim

More information

Monte Carlo Optimization to Solve a Two-Dimensional Inverse Heat Conduction Problem

Monte Carlo Optimization to Solve a Two-Dimensional Inverse Heat Conduction Problem Australia Joural of Basic Applied Scieces, 5(): 097-05, 0 ISSN 99-878 Mote Carlo Optimizatio to Solve a Two-Dimesioal Iverse Heat Coductio Problem M Ebrahimi Departmet of Mathematics, Karaj Brach, Islamic

More information

Entropy and Ergodic Theory Lecture 5: Joint typicality and conditional AEP

Entropy and Ergodic Theory Lecture 5: Joint typicality and conditional AEP Etropy ad Ergodic Theory Lecture 5: Joit typicality ad coditioal AEP 1 Notatio: from RVs back to distributios Let (Ω, F, P) be a probability space, ad let X ad Y be A- ad B-valued discrete RVs, respectively.

More information

DIGITAL SIGNAL PROCESSING LECTURE 3

DIGITAL SIGNAL PROCESSING LECTURE 3 DIGITAL SIGNAL PROCESSING LECTURE 3 Fall 2 2K8-5 th Semester Tahir Muhammad tmuhammad_7@yahoo.com Cotet ad Figures are from Discrete-Time Sigal Processig, 2e by Oppeheim, Shafer, ad Buc, 999-2 Pretice

More information

The DOA Estimation of Multiple Signals based on Weighting MUSIC Algorithm

The DOA Estimation of Multiple Signals based on Weighting MUSIC Algorithm , pp.10-106 http://dx.doi.org/10.1457/astl.016.137.19 The DOA Estimatio of ultiple Sigals based o Weightig USIC Algorithm Chagga Shu a, Yumi Liu State Key Laboratory of IPOC, Beijig Uiversity of Posts

More information

Chapter 22 Developing Efficient Algorithms

Chapter 22 Developing Efficient Algorithms Chapter Developig Efficiet Algorithms 1 Executig Time Suppose two algorithms perform the same task such as search (liear search vs. biary search). Which oe is better? Oe possible approach to aswer this

More information

A New Bound between Higher Order Nonlinearity and Algebraic Immunity

A New Bound between Higher Order Nonlinearity and Algebraic Immunity Available olie at wwwsciecedirectcom Procedia Egieerig 9 (01) 788 79 01 Iteratioal Workshop o Iformatio ad Electroics Egieerig (IWIEE) A New Boud betwee Higher Order Noliearity ad Algebraic Immuity Xueyig

More information

POWER SERIES SOLUTION OF FIRST ORDER MATRIX DIFFERENTIAL EQUATIONS

POWER SERIES SOLUTION OF FIRST ORDER MATRIX DIFFERENTIAL EQUATIONS Joural of Applied Mathematics ad Computatioal Mechaics 4 3(3) 3-8 POWER SERIES SOLUION OF FIRS ORDER MARIX DIFFERENIAL EQUAIONS Staisław Kukla Izabela Zamorska Istitute of Mathematics Czestochowa Uiversity

More information

On Extracting Properties of Lie Groups from Their Lie Algebras

On Extracting Properties of Lie Groups from Their Lie Algebras America Joural of Computatioal ad Applied Mathematics 216, 6(5): 182-186 DOI: 1.5923/j.ajcam.21665.2 O Extractig Properties of Lie Groups from Their Lie Algebras M-Alami A. H. Ahmed 1,2 1 Departmet of

More information

Analog and Digital Signals. Introduction to Digital Signal Processing. Discrete-time Sinusoids. Analog and Digital Signals

Analog and Digital Signals. Introduction to Digital Signal Processing. Discrete-time Sinusoids. Analog and Digital Signals Itroductio to Digital Sigal Processig Chapter : Itroductio Aalog ad Digital Sigals aalog = cotiuous-time cotiuous amplitude digital = discrete-time discrete amplitude cotiuous amplitude discrete amplitude

More information

Using the IML Procedure to Examine the Efficacy of a New Control Charting Technique

Using the IML Procedure to Examine the Efficacy of a New Control Charting Technique Paper 2894-2018 Usig the IML Procedure to Examie the Efficacy of a New Cotrol Chartig Techique Austi Brow, M.S., Uiversity of Norther Colorado; Bryce Whitehead, M.S., Uiversity of Norther Colorado ABSTRACT

More information