A HIERARCHICAL SIMULATION FRAMEWORK FOR APPLICATION DEVELOPMENT ON SYSTEM-ON-CHIP ARCHITECTURES. Vaibhav Mathur and Viktor K.

Size: px
Start display at page:

Download "A HIERARCHICAL SIMULATION FRAMEWORK FOR APPLICATION DEVELOPMENT ON SYSTEM-ON-CHIP ARCHITECTURES. Vaibhav Mathur and Viktor K."

Transcription

1 A HIERARCHICAL SIMULATION FRAMEWORK FOR APPLICATION DEVELOPMENT ON SYSTEM-ON-CHIP ARCHITECTURES Vabhav Mathur and Vktor K. Prasanna Department of EE-Systems Unversty of Southern Calforna Los Angeles, CA fvabhav, ABSTRACT We propose a herarchcal smulaton methodology to assst applcaton development on System-on-Chp archtectures. Herarchcal smulaton nvolves smulaton of a SoC based system at dfferent levels of abstracton. Thus, t enables a system desgner to explot smulaton speed vs. accuracy of results trade-offs. Vertcal smulaton s a specal case of herarchcal smulaton, where a feedback mechansm between the dfferent smulaton levels helps n nterpretng the results of stand-alone smulatons n the system-wde context. The paper presents an approach to perform vertcal smulaton of a class of applcatons under a smplfed scenaro. I. INTRODUCTION Performance requrements under ubqutous computng, and convergence of communcaton and computng technologes have resulted n the emergence of System-on- Chp (SoC) archtectures. Applcaton development on such archtectures nvolves a delcate balance between hgh performance requrements and constrants on area, power, etc. State-of-the-art desgn tools and methodologes are not adequate to manage the desgn complexty of SoCs. Current desgn processes are based on ndependent desgn flow for each archtecture component and have not co-evolved wth changng system desgns and requrements. Programmng models and desgn tools for each component are ndependently utlzed to map an applcaton, and system ntegraton s performed later. Systemwde performance analyss s typcally a manual process. It nvolves the use of component specfc smulators n solaton, and s tedous snce each smulator has a dfferent nput/output nterface. Ths approach results n suboptmal desgn because mult-objectve optmzaton requres exhaustve traversal of a large desgn space. Several envronments [9], [15] have emerged that concentrate more on co-smulaton and system synthess than on hgh level system desgn. Desgn problems (e.g. mproper applcaton-to-archtecture mappng, nsuffcent resources to meet the performance requrements, etc.) detected durng such co-smulaton leads to tedous and tme consumng redesgn of the system [6]. Ths work s supported by the US DARPA Power Aware Computng and Communcaton Program under contract F33615-C montored by Wrght Patterson Ar Force Base. Analog I/O Programmable Processor Customzed Logc Memory Subsystem Hgh Speed Interconnect DMA Subsystem Confgurable Logc Fg. 1. A typcal SoC archtecture The above scenaro necesstates a system level desgn approach. A unfed smulaton envronment s needed that provdes performance estmates (latency, throughput, average power, etc.), for a gven applcaton-to-archtecture mappng, at desred detal and cost (tme) of smulaton. Ths wll enable rapd evaluaton of performance tradeoffs for alternate mappngs at a hgh level and early n the desgn cycle wthout performng tme-ntensve low-level smulatons. Ths paper outlnes a herarchcal smulaton methodology for ths purpose. The work descrbed n ths paper s part of the MILAN 1 project, whch s a collaboratve research effort between the Unversty of Southern Calforna (USC) and the Insttute for Software Integrated Systems (ISIS) at Vanderblt Unversty. MILAN s a Model based Integrated smulaton framework for embedded system desgn and optmzaton [2]. The desgner formally models the target applcaton, underlyng hardware, and constrants (latency, throughput, power, etc.) through a graphcal nterface provded by MILAN. The nformaton s stored n a model database that can be accessed through a smple Component Object Model (COM) [13] nterface. The model nformaton s translated nto sutable nput formats requred by the ntegrated smulators. Thus, MILAN has the capablty to drve multple smulators wth dfferent nput/output formats from a sngle system specfcaton. Herarchcal smulaton means smulatng a desgn at dfferent levels of abstracton n terms of both the structure and the behavor of the underlyng component(s). Vertcal smulaton s a specal case of herarchcal smula- 1 mlan (hnd): meetng, unfcaton.

2 System Development Process: Defne Modelng Paradgm Tools and Components: Creates Develop Model Bulder(s) Model Bulder Use End-User s Process: Edt Models Develop Model Interpreter Creates Model Interpreter Models Use Buld System (Automatcally) Executable System Develop R-T Support Creates Run-tme Support Execute System Fg. 2. Model Integrated Computng ton, where a desgner s nterested not only n smulatng a system sub-component (a task-resource mappng) at multple levels of granularty, but also n meanngful nterpretaton of low-level smulaton results from a systemwde perspectve. For example, the throughput of one system sub-component mght affect the rate of avalablty of data to a data dependent system sub-component. Also, the computaton of a partcular task mght depend on nput data values, that can sometmes only be determned by smulatng real-world system operaton va realstc data sets. Vertcal smulaton s useful when (a) hgh level end-to-end performance estmates are to be further refned through low-level smulaton, and (b) a partcular system sub-component s to be smulated at a fne granularty, takng nto account system-wde effects such as those mentoned above. A hgh level estmaton methodology and a feedback mechansm, explaned n detal n Sectons II and III, helps n achevng both these objectves. Vertcal smulaton wll typcally be useful n the later phase of the desgn cycle, when hgh-level, system-wde performance estmates for a partcular mappng are to be refned through low-level smulatons. Hgh-level smulatons are coarse-graned approxmatons whch yeld very rapd results wth possble compromse n accuracy. Lowlevel smulatons typcally are hghly accurate but are tme consumng. Herarchcal smulaton wth nterpretaton of smulaton results allows the user to explot the trade-offs between smulaton speed and accuracy of results. MILAN focuses prmarly on ntegraton of smulators for Instructon Set Archtecture (ISA) based processors (e.g. RISC, DSP), applcaton-specfc cores, confgurable logc (e.g. FPGA), memores, and nterconnects. Numerous commercal platforms targeted towards communcaton and networkng applcatons, have been ntroduced lately. A typcal platform s shown n Fgure 1. The Unversal Mcrosystem (UMS) [8], the Reconfgurable Communcatons Processor [7], the Jazz PSA [11], and the Platform FPGA [16], are a few examples. These platforms nclude programmable processor(s), confgurable logc, customzed logc, embedded memores, hgh speed nterconnect, and analog I/O components on the same chp. MI- LAN ntegrates smulators for dgtal components only; smulators for analog components are not consdered. The paper s organzed as follows. Secton II descrbes the MILAN project and the underlyng Model Integrated Computng (MIC) [18] desgn approach. Secton III defnes herarchcal and vertcal smulaton concepts. Secton IV dscusses a prototype mplementaton of vertcal smulaton for a class of applcatons. Secton V has concludng remarks. II. THE MILAN PROJECT The focus of the MILAN project [2] s on developng formal modelng paradgms that wll enable smulator ntegraton and effcent applcaton-to-archtecture mappng through automatc desgn space exploraton. MI- LAN adopts Model Integrated Computng (MIC) [18] as the core desgn technology (Fgure 2). MIC s especally valuable for the desgn of computer-based systems wth strong nterdependence between the hardware and the software components. By formally modelng all aspects (applcaton, resource, behavor, constrants, etc.) of a system and usng well-defned rules to generate new systems or manage exstng ones, t s possble to avod the errors that arse when requrements change and the system has to be redesgned or re-mplemented. Whle the ntal modelng effort mght be costly compared to ad hoc approaches, the benefts are apparent for a system that evolves over tme. An envronment that supports the MIC allows the desgner to create doman-specfc models at the requred level of abstracton, valdate these models, and perform varous computatonal transformatons on them (Fgure 2).

3 Functonal Smulator Power Smulator Applcaton Models Generc Modelng Envronment (GME) Constrants Desgn-space Exploraton and Prunng DESIGN System Smulaton SYSTEM Resource Models Fg. 3. MILAN Archtecture Hgh-Level Estmator Performance Smulator Model Interpreters A metamodel s a formal descrpton of the modelng envronment s model constructon semantcs,.e. the modelng paradgm. Syntactcally and semantcally correct models are created usng modelng paradgms. The metamodel defnes the syntactcal buldng blocks avalable to the desgner for descrbng (nstantatng) the system. It also defnes the doman-specfc composblty rules and the constrants that enforce some of the semantcs n the system specfcaton. Model nterpreters are the software components that translate the models for use n the MILAN executon envronment. A model database s a key component that stores the models and the translated nformaton n a canoncal form that provdes a common representaton for the nformaton that are used n drvng varous smulators. The model database also stores the smulaton results. The Generc Modelng Envronment (GME) [10] s a confgurable graphcal tool sute supportng MIC. The confguraton of the envronment to support domanspecfc modelng s performed n a formal manner through the use of metamodels. The metamodelng language s the UML class dagram notaton [5]. Well-formedness rules that are also part of the metamodels are specfed usng the Object Constrant Language (OCL). These constrants, along wth the syntactcal rules of the doman language, are enforced by the automatcally generated target envronment. MILAN explots the MIC technology to confgure an envronment talored for embedded system desgn, evaluaton, and optmzaton. The framework ncorporates power as an mportant desgn metrc. Power estmaton and optmzaton s supported through ntegraton of exstng component specfc power smulators, based on system-wde power models. Fgure 3 shows the archtecture of MILAN and also depcts the system desgn flow from the users perspec- tve. The graphcal nterface s provded by GME confgured to support the modelng paradgms developed for MI- LAN. Semantc nformaton n the modelng paradgms s captured through metamodels. The metamodels confgure the GME for creatng doman-specfc models. Followng models for the desgn of embedded systems based on SoC archtectures are currently supported n MILAN. Resource models [3] descrbe avalable hardware components and ther nterconnectvty n a herarchcal block dagram-lke notaton. Applcaton models are based on a herarchcal sgnal flow representaton wth mportant extensons. Most notably, the modelng language allows for the specfcaton of explct desgn or mplementaton alternatves of any component. Ths enables modelng of the entre desgn space of the applcaton as opposed to a pont soluton. To manage ths desgn space, applcaton requrements, resource constrants, and other specfcatons are captured explctly through OCL, n the metamodels. Performance modelng of SoC archtectures nvolves characterzng desred performance metrcs of a gven mappng n terms of archtecture parameters. The performance model leverages from pror USC work n hgh level performance models for tradtonal and reconfgurable archtectures [1][14]. The communcaton model provdes a common formalsm to enable nter-operablty of smulators that represent the same nformaton n dfferent formats. The desgn-space exploraton and prunng tool takes the potentally very large desgn space and apples the constrants usng a symbolc constrant satsfacton technque to fnd the set of solutons that satsfy all the constrants. The goal of desgn space exploraton s to dentfy a small number of vald canddate desgns. MILAN supports an estmaton methodology that ntegrates varous component-specfc performance models (power, tme) and enables system-wde performance evaluaton. The methodology s based on a hgh level systemwde performance model to evaluate complex applcaton to archtecture mappng choces and dfferent schedules of executon. Although, deally a low-level (detaled) system wde performance model can be defned, t s mpractcal snce dverse archtecture features (reconfgurable vs. customzed logc, varable vs. statc parameters etc.) cannot be captured unformly. Moreover, smulatons based on t wll be cost (tme) ntensve. The desgn of a Hgh Level Performance Estmator (HPerE) based on a hgh level performance model s motvated by these ssues. MILAN supports dfferent classes of smulators, besdes the HPerE. Functonal smulators (e.g., MATLAB or SystemC), verfy the functonalty of the applcaton. Several low-level smulators (power, performance) wll be supported n future. The HPerE along wth the low-level smulators ntegrated n MILAN facltates mult-level smulaton, whch explots the trade-off between the accuracy of results and the smulaton speed.

4 III. HIERARCHICAL AND VERTICAL SIMULATION MILAN has the followng capabltes to assst applcaton development under a unfed smulaton envronment. ffl Smulator Integraton: Integratng multple component specfc smulators and drvng these smulators wth dfferent nput/output formats usng a sngle system specfcaton s defned as Smulator Integraton. The system s specfed n terms of applcaton and resource models, and end-to-end performance requrements. The models are stored n the model database that enables seamless smulator ntegraton by provdng a common nformaton repostory. The model nterpreters translate ths nformaton nto the syntax and the semantcs of the nputs requred to drve the smulators. ffl Herarchcal smulator ntegraton: Herarchcal refers to multple levels of abstracton of the models descrbng the system. Herarchcal smulaton mples smulatng the system or a system sub-component at dfferent levels of abstracton and mplementaton. Thus, provdng herarchcal smulaton capablty for a partcular task-to-resource mappng means provdng the user wth a choce of smulators for that resource, at dfferent levels of granularty (assumng they are ntegrated nto the framework). Smulator ntegraton of component-specfc or system-wde hgh level and low level smulators s termed as herarchcal smulator ntegraton. Hgh-level smulators are typcally based on a few key parameters of the system (or system sub-component) and are desgned to provde rapd performance estmates, possbly at the expense of accuracy. Low-level smulators (such as cycle-accurate smulators) are hghly accurate but can requre an order of magntude more smulaton tme and a much larger set of nput parameters to be specfed, compared to hgh-level smulators. The herarchcal, block-dagram lke specfcaton of the resources themselves also provdes dfferent levels of abstracton correspondng to the dfferent granulartes of resource representaton requred by hgh-level and low-level smulators. ffl Vertcal smulaton: Herarchcal smulator ntegraton enables the smulaton of a system sub-component (a task mapped onto a compute resource) at dfferent levels of granularty, usng any of the ntegrated smulators for that component, thereby allowng the desgner to explot the smulaton tme vs. accuracy of results trade-off. There are two scenaros when stand-alone, mult-granular smulaton s not useful: (a) when the desgner s nterested n refnng hgh-level estmates of a partcular mappng, provded by HPerE, and (b) when the desgner wants detaled and realstc statstcs about a partcular system sub-component. A meanngful nterpretaton of the results of herarchcal (mult-granular) smulaton, n the system-wde context s defned as vertcal smulaton. Vertcal smulaton uses the herarchcal smulator ntegraton capablty and provdes Functonal Smulaton User System Specfcaton DATABASE Model Informaton Stmulus Vertcal Smulaton Low level Smulators HPerE Model Interpreters Fg. 4. Herarchcal Smulaton Framework drvng smulators feedback a mechansm to refne a system sub-component or the endto-end performance estmates of a canddate mappng. The HPerE system-wde performance estmaton methodology does not rely on cycle-accurate smulaton of a gven applcaton on the target resource at the tme HPerE s nvoked. The HPerE requres pre-characterzed costs from expermental results, vendor-suppled nformaton for lbrary components, or any other offlne method. In the later stages of the desgn cycle, smulaton tme wll be of less concern because the desgn space would have reduced to a smaller, manageable set of mappngs that satsfy broad system requrements. In the frst desgn scenaro as mentoned above, the desgner wll be nterested n refnng the HPerE estmates for a gven mappng, by runnng low-level smulatons for selected components. The model database not only stores the models but also the data values processed by the system sub-component, and the results of smulaton. One of the crucal nformaton requred for nvokng any smulator s the ncomng data set(s), whch form a stmulus to the system sub-component (a task-tocompute resource mappng). The stmulus s obtaned ether through functonal smulaton of the applcaton or by usng data generator scrpts for the predecessor tasks n the path of applcaton executon. These methods are explaned n detal n Secton IV. Snce the smulaton results are also stored n the database, smulatng a component at

5 low-level mples updatng the performance cost assocated wth t. By nvokng the approprate low-level smulator, updatng the performance costs, and then nvokng HPerE, refned estmates can be obtaned (for the frst scenaro). Ths bottom-up feedback needs model nterpreters that can sutably nterpret the low-level smulaton results and update the model database. The second scenaro s much more challengng because t requres a top-down feedback from the HPerE to enable realstc low-level smulaton, as aganst the bottomup feedback from the low-level smulaton results for refnng HPerE estmates. A realstc smulaton s one that provdes performance statstcs closest to those exhbted by the fnal system when t s deployed n the real world. Two of the mportant consderatons n arrvng at such realstc low-level smulaton results are the avalablty of realstc data values that wll be processed by that task, and modelng of system-wde effects such as nput/output delays due to samplng rates of sensors, constrants on buffer sze, etc. Although the HPerE and low-level smulaton results are both stored n the model database, a common respostory by tself s not suffcent to help the user get a realstc set of performance statstcs for a partcular task, wth system-wde effects taken nto consderaton. Vertcal smulaton technques are requred to meanngfully nterpret low-level smulaton results wth top-down feedback from the HPerE. The next secton addresses ths ssue n more detal. There s another aspect to ntegratng smulators that s outsde the scope of the MILAN project. When more than one component-specfc low-level smulators are nvoked smultaneously and system-wde performance s evaluated through real-tme nteracton among the dfferent smulators, the capablty s termed horzontal smulaton. Although horzontal smulaton, f mplemented correctly, can accurately smulate the actual nteracton among subcomponents, concurrent executon of component-specfc smulators poses multtude of challenges. For example, most wdely-used smulators are not desgned to nteract at run-tme, wth other systems. Modfyng such smulators can prove to be tme-consumng and not always feasble. Specfyng a common API for smulators to conform to, s an equally challengng task. Fnally, there are smulaton speed (tme) ssues that can arse when smulators have to synchronze, say, at every cycle. IV. VERTICAL SIMULATION: ASIMPLE IMPLEMENTATION Usng the MILAN modelng paradgms, the system desgner specfes the target applcaton, underlyng resources, the task-to-resource mappng, a schedule of executon of the tasks, and performance constrants. The nformaton s stored n the model database, and model nterpreters extract ths nformaton and translate t nto sutable formats for drvng the ntegrated smulators. Herarchcal smulaton s facltated through ths mechansm, as explaned n the prevous secton. Ths secton outlnes vertcal smulaton mplementaton for a smplfed scenaro. Followng assumptons are made. ffl Applcaton s modeled as a Drected Acyclc Graph (DAG). Each vertex n the DAG represents a task and edges between the vertces represent the data transfer between the tasks. Every task n the DAG s atomc. Ths means that a task cannot be further decomposed nto subtasks. If the task s dependent on data from multple tasks, we assume that the task s actvated (smulated) when the data from all ts predecessor tasks (the stmulus) n the DAG s avalable. ffl Currently, the data processng rate by all the tasks s assumed to be same and constant durng the entre executon of the applcaton. There s no change n the rate of processng dependent on a control data. ffl Resources consst of compute resources (RISC core, FPGA, etc.), storage elements (memores), and communcaton channels (buses). Explct communcaton through the channels occurs between the compute resources. ffl The task-to-resource mappng and the schedule of executon of the tasks on the compute resources together defne an executon model for the applcaton. Only one task s executed on a compute resource at any tme. There s no resource sharng wth others tasks durng the executon.e., the tasks do not compete for resources such as compute cycles, memory etc. The precedng (and succeedng) tasks n the schedule can however be executed on the same resource. Snce there s no concurrent executon of the tasks on the same resource, the results of stand-alone smulaton (that typcally does not model concurrency and resource sharng) accurately represent the computaton cost of the task beng smulated. A lnear task graph wth only one task mapped onto a compute resource at a tme, trvally satsfes the assumptons under the applcaton and the executon model. The sequental data dependence between the tasks satsfes the requrement that the stmulus (consstng of just one data value) s avalable before the task s ntated. For vertcal smulaton, the user selects a partcular task and an assocated low-level smulator for the resource onto whch that task s mapped. Invocaton of a smulator requres the followng nputs: (a) The stmulus (the ncomng data set to the task) (b) the smulator confguraton parameters, and (c) an approprate mplementaton of the task (e.g. C code for a RISC smulator or a VHDL code for an FPGA smulator). These three nputs are provded by the model nterpreter. The stmulus S to a task denoted by vertex V s modeled by two vectors: S =[fd 1 ;d 2 ; :::; d g; k ft 1 ;t 2 ; g], :::t k where d represents nput data value, t represents the arrval tme of the value, and k s the n-degree of V. The

6 V V V V Database Stmulus, task code, confguraton parameters (a) Step I: Input DAG G d 2 d 1 d1 d2 (b) Step II: Back traversal to the nput vertex wth reversed edges Low level smulator S vs = {d 1, d 2 } (c) Step III: Functonal Smulaton computaton statstcs (d) Step V: Low level smulatons Fg. 5. Steps for vertcal smulaton rate of arrval and the data values are crucal for realstc smulaton of the task. Our smple vertcal smulaton technque focuses on task latency and not throughput,.e. the arrval rate of data at a node s not currently modeled. For determnng the stmulus, the followng method s used (see Fgure 5). The task to be vertcally smulated, and the nput vertex V for the entre applcaton are marked n the applcaton DAG G. To determne all data flows that lead from V to, the drecton of the edges s reversed and the graph s traversed startng from. All the vertces on all the paths leadng from to the nput vertex V are determned. Functonal smulaton s then performed for all these vertces (tasks) accordng to the dependences specfed by the orgnal graph G, and the data stmulus for s obtaned. Another method for obtanng the stmulus s requrng the user to assocate a scrpt wth every task. Ths scrpt generates dummy output that represents the actual output, had the task been executed wth real nput data. If such scrpts are provded, obtanng the data stmulus for a specfc task wll nvolve executng the dummy data generators only for ts mmedately precedng tasks. Functonal smulaton s stll preferable, because the computaton cost of a task s most accurately obtaned through smulaton that uses data values generated durng the actual applcaton executon. For example, consder an mage processng applcaton. Computaton of the correlaton coeffcents for an mage vector depends on the ntensty of the pxels, whch n turn depends on the relatve dstance on the RGB color scale. Generatng dummy vectors that represent actual mages seen by the system n the real world can be dffcult or even mpossble. Whle the stmulus obtaned through functonal smulaton and generator scrpts s stored n the model database, some typcal data values generated durng pror applcaton development, and tested for the worst or the best case performance satsfacton can as well be stored. These can be reused for the current applcaton development process as stmulus for low-level smulatons. Once the stmulus s obtaned, the approprate low-level smulator s nvoked. The results from low-level smulaton need to be sutably nterpreted and used to update the computaton costs assocated wth the task n the model database. Ths s a bottom-up feedback mechansm, where results from low-level smulaton are used to obtan more accurate system-wde performance estmates through HPerE. Stand-alone smulaton of a task, however, does not model the system-wde effects on task executon, such as nput/output delays due to samplng rates of sensors, constrants on buffer sze, etc. The smulaton also does not capture the state of the processng element between task executons. For example, the cache contents after a task completes executon affect the ntal memory access costs (compulsory msses) for the next task that executes on the same resource. These access costs wll be avalable through HPerE and wll also be stored as a stmulus n the database. However, ths stmulus s requred wth the output of the low-level smulatons. HPerE provdes such top-down feedback necessary for system-wde nterpretaton of the low-level smulaton results. Theestmates of system-wde effects provded by HPerE, such as nput/output data delays, are used n conjuncton wth stand-alone smulaton results to arrve at more accurate performance statstcs for the system sub-component. The ssues n mplementng vertcal smulaton for a general scenaro are dscussed n Secton V. The vertcal smulaton outlned above s beng mplemented n MILAN. The MILAN applcaton modelng paradgm allows graphcal representaton of the target applcaton as a DAG, wth task mplementatons specfed for each node of the DAG (C source code, HDL mplementatons, etc.). The current verson of the resource modelng paradgm s capable of representng unprocessor archtectures as modeled by the SmpleScalar smulator [17], a prototype ntegraton of whch has been completed for MI-

7 LAN. Currently, for specfyng task-to-resource mappng, a task node n the applcaton graph s explctly assocated wth the underlyng resource label. MATLAB has been ntegrated nto MILAN for functonal smulaton. These capabltes of MILAN enable a user to functonally smulate the entre applcaton, and perform low-level smulaton for a selected task-resource par. Vertcal smulaton, as descrbed prevously, requres addtonal capabltes that are beng mplemented. These nclude graph traversal from the specfed node n the drecton of the edges, functonal smulaton of an applcaton sub-graph, desgn and mplementaton of model nterpreters for bottom-up feedback, etc. A prelmnary desgn of HPerE has been completed, and mplementaton s under progress. Vertcal smulaton for applcaton development provdes the capablty to perform mult-level smulaton for a partcular applcaton-to-resource mappng. The desgner can evaluate hardware vs. software speedup for a task and can change the mappng for that task only, or for the entre system f desred. An FFT algorthm mplemented as Decmaton n Tme and Frequency (DITF) s less computatonally ntensve (n terms of floatng pont operatons) as compared to DIT or DIF radx-2 mplementatons [4]. A desgner can evaluate the end-to-end performance enhancements usng such alternate algorthmc mplementatons. V. CONCLUDING REMARKS Ths paper dscussed the concept and need for herarchcal and vertcal smulaton, and a smple mplementaton n the context of MILAN. Some mportant ssues that are applcable n a more general scenaro reman to be addressed. A few of the mportant characterstcs of the general scenaro are: ffl Tasks are executed concurrently on the same resource. ffl The latency and throughput of a task change dependng on control data. Ths n turn, affects the rate of avalablty of data to the data dependent tasks n the applcaton executon. ffl Communcaton between compute resources s through dfferent mechansms, such as global memory. Ths mpacts the end-to-end latency, power consumpton, etc. We are currently enhancng HPerE to address the ssues n the general scenaro. The feedback mechansm (bottom-up and top-down) as dscussed n the paper wll be mplemented and enhanced. Ths wll nvolve wrtng approprate model nterpreters. In future, the refnement of the hgh level model parameters themselves wll be performed, based on low-level smulaton results. We envson that our framework wll address the system level desgn challenge facng the SoC communty by provdng an accurate mechansm to evaluate nteractons between SoC components at varous levels of granularty, by explotng the smulaton tme vs. accuracy of results trade-off. VI. ACKNOWLEDGMENTS We would lke to thank Amol Baksh (USC), Sumt Mohanty (USC), and Akos Ledecz (ISIS). Amol provded nvaluable feedback on smulator ntegraton, and herarchcal smulaton ntegraton aspects of MILAN (the subject of ths paper). Sumt provded nputs on HPerE, and Akos provded nputs on the MILAN archtecture. REFERENCES [1] Algorthms for Data Intensve Applcatons on Intellgent and Smart Memores (ADVISOR), Unv. of Southern Calforna. [2] A. Baksh, V. K. Prasanna, A. Ledecz, et al., MILAN: A Model Based Integrated Smulaton Framework for Desgn of Embedded Systems, ACM SIGPLAN 2001 Workshop on Languages, Complers, and Tools for Embedded Systems (LCTES 2001), Utah, June [3] A. Baksh and V. K. Prasanna, Abstract Resource Representatons for Custom Desgn of System-on-Chp Archtectures, submtted to IFIP VLSI-SOC 2001, Montpeller, France, December [4] M. Balducc, A. Ganapathraju et al., Benchmarkng of FFT Algorthms, IEEE Southeastcon 97, Engneerng New Century, Proceedngs, pp [5] G. Booch et al., The Unfed Modelng Language User Gude, Addson-Wesley Pub Co., [6] H. Chang et al., Survvng the SOC Revoluton A Gude to Platform-Based Desgn, Kluwer Academc Publsher, Massachusetts, USA, [7] Chameleon Systems Inc., [8] Cradle Technologes, [9] Synopsys Eagle, ds.html. [10] Generc Modelng Envronment, [11] Improv Systems Inc., [12] The MILAN Project, [13] Mcrosoft Component Object Model, [14] Models, Algorthms and Archtectures for Reconfgurable Computng Project, [15] CoWare N2C, [16] Xlnx Inc., [17] SmpleScalar Tool Set, mscalar/smplescalar.html. [18] J. Sztpanovts, G. Karsa, Model-Integrated Computng, IEEE Computer, Aprl, 1997.

Model Integrated Computing: A Framework for Creating Domain Specific Design Environments

Model Integrated Computing: A Framework for Creating Domain Specific Design Environments Model Integrated Computng: A Framework for Creatng Doman Specfc Desgn Envronments James R. DAVIS Vanderblt Unversty, Insttute for Software Integrated Systems Nashvlle, TN 37203, USA ABSTRACT Model Integrated

More information

Multigranular Simulation of Heterogeneous Embedded Systems

Multigranular Simulation of Heterogeneous Embedded Systems Multgranular Smulaton of Heterogeneous Embedded Systems Adtya Agrawal Insttute for Software Integrated Systems Vanderblt Unversty Nashvlle, TN - 37235 1 615 343 7567 adtya.agrawal@vanderblt.edu Akos Ledecz

More information

A Binarization Algorithm specialized on Document Images and Photos

A Binarization Algorithm specialized on Document Images and Photos A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a

More information

An Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

More information

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009.

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009. Farrukh Jabeen Algorthms 51 Assgnment #2 Due Date: June 15, 29. Assgnment # 2 Chapter 3 Dscrete Fourer Transforms Implement the FFT for the DFT. Descrbed n sectons 3.1 and 3.2. Delverables: 1. Concse descrpton

More information

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz Compler Desgn Sprng 2014 Regster Allocaton Sample Exercses and Solutons Prof. Pedro C. Dnz USC / Informaton Scences Insttute 4676 Admralty Way, Sute 1001 Marna del Rey, Calforna 90292 pedro@s.edu Regster

More information

Simulation Based Analysis of FAST TCP using OMNET++

Simulation Based Analysis of FAST TCP using OMNET++ Smulaton Based Analyss of FAST TCP usng OMNET++ Umar ul Hassan 04030038@lums.edu.pk Md Term Report CS678 Topcs n Internet Research Sprng, 2006 Introducton Internet traffc s doublng roughly every 3 months

More information

The Codesign Challenge

The Codesign Challenge ECE 4530 Codesgn Challenge Fall 2007 Hardware/Software Codesgn The Codesgn Challenge Objectves In the codesgn challenge, your task s to accelerate a gven software reference mplementaton as fast as possble.

More information

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty

More information

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth

More information

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,

More information

Petri Net Based Software Dependability Engineering

Petri Net Based Software Dependability Engineering Proc. RELECTRONIC 95, Budapest, pp. 181-186; October 1995 Petr Net Based Software Dependablty Engneerng Monka Hener Brandenburg Unversty of Technology Cottbus Computer Scence Insttute Postbox 101344 D-03013

More information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism for Nested Loops with Non-uniform and Flow Dependences Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr

More information

Some material adapted from Mohamed Younis, UMBC CMSC 611 Spr 2003 course slides Some material adapted from Hennessy & Patterson / 2003 Elsevier

Some material adapted from Mohamed Younis, UMBC CMSC 611 Spr 2003 course slides Some material adapted from Hennessy & Patterson / 2003 Elsevier Some materal adapted from Mohamed Youns, UMBC CMSC 611 Spr 2003 course sldes Some materal adapted from Hennessy & Patterson / 2003 Elsever Scence Performance = 1 Executon tme Speedup = Performance (B)

More information

Cluster Analysis of Electrical Behavior

Cluster Analysis of Electrical Behavior Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School

More information

Verification by testing

Verification by testing Real-Tme Systems Specfcaton Implementaton System models Executon-tme analyss Verfcaton Verfcaton by testng Dad? How do they know how much weght a brdge can handle? They drve bgger and bgger trucks over

More information

Assembler. Building a Modern Computer From First Principles.

Assembler. Building a Modern Computer From First Principles. Assembler Buldng a Modern Computer From Frst Prncples www.nand2tetrs.org Elements of Computng Systems, Nsan & Schocken, MIT Press, www.nand2tetrs.org, Chapter 6: Assembler slde Where we are at: Human Thought

More information

AADL : about scheduling analysis

AADL : about scheduling analysis AADL : about schedulng analyss Schedulng analyss, what s t? Embedded real-tme crtcal systems have temporal constrants to meet (e.g. deadlne). Many systems are bult wth operatng systems provdng multtaskng

More information

TN348: Openlab Module - Colocalization

TN348: Openlab Module - Colocalization TN348: Openlab Module - Colocalzaton Topc The Colocalzaton module provdes the faclty to vsualze and quantfy colocalzaton between pars of mages. The Colocalzaton wndow contans a prevew of the two mages

More information

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique //00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy

More information

Topology Design using LS-TaSC Version 2 and LS-DYNA

Topology Design using LS-TaSC Version 2 and LS-DYNA Topology Desgn usng LS-TaSC Verson 2 and LS-DYNA Wllem Roux Lvermore Software Technology Corporaton, Lvermore, CA, USA Abstract Ths paper gves an overvew of LS-TaSC verson 2, a topology optmzaton tool

More information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,

More information

PRÉSENTATIONS DE PROJETS

PRÉSENTATIONS DE PROJETS PRÉSENTATIONS DE PROJETS Rex Onlne (V. Atanasu) What s Rex? Rex s an onlne browser for collectons of wrtten documents [1]. Asde ths core functon t has however many other applcatons that make t nterestng

More information

Mathematics 256 a course in differential equations for engineering students

Mathematics 256 a course in differential equations for engineering students Mathematcs 56 a course n dfferental equatons for engneerng students Chapter 5. More effcent methods of numercal soluton Euler s method s qute neffcent. Because the error s essentally proportonal to the

More information

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 A mathematcal programmng approach to the analyss, desgn and

More information

Real-Time Systems. Real-Time Systems. Verification by testing. Verification by testing

Real-Time Systems. Real-Time Systems. Verification by testing. Verification by testing EDA222/DIT161 Real-Tme Systems, Chalmers/GU, 2014/2015 Lecture #8 Real-Tme Systems Real-Tme Systems Lecture #8 Specfcaton Professor Jan Jonsson Implementaton System models Executon-tme analyss Department

More information

Real-time Motion Capture System Using One Video Camera Based on Color and Edge Distribution

Real-time Motion Capture System Using One Video Camera Based on Color and Edge Distribution Real-tme Moton Capture System Usng One Vdeo Camera Based on Color and Edge Dstrbuton YOSHIAKI AKAZAWA, YOSHIHIRO OKADA, AND KOICHI NIIJIMA Graduate School of Informaton Scence and Electrcal Engneerng,

More information

an assocated logc allows the proof of safety and lveness propertes. The Unty model nvolves on the one hand a programmng language and, on the other han

an assocated logc allows the proof of safety and lveness propertes. The Unty model nvolves on the one hand a programmng language and, on the other han UNITY as a Tool for Desgn and Valdaton of a Data Replcaton System Phlppe Quennec Gerard Padou CENA IRIT-ENSEEIHT y Nnth Internatonal Conference on Systems Engneerng Unversty of Nevada, Las Vegas { 14-16

More information

Module Management Tool in Software Development Organizations

Module Management Tool in Software Development Organizations Journal of Computer Scence (5): 8-, 7 ISSN 59-66 7 Scence Publcatons Management Tool n Software Development Organzatons Ahmad A. Al-Rababah and Mohammad A. Al-Rababah Faculty of IT, Al-Ahlyyah Amman Unversty,

More information

Evaluation of an Enhanced Scheme for High-level Nested Network Mobility

Evaluation of an Enhanced Scheme for High-level Nested Network Mobility IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.15 No.10, October 2015 1 Evaluaton of an Enhanced Scheme for Hgh-level Nested Network Moblty Mohammed Babker Al Mohammed, Asha Hassan.

More information

Configuration Management in Multi-Context Reconfigurable Systems for Simultaneous Performance and Power Optimizations*

Configuration Management in Multi-Context Reconfigurable Systems for Simultaneous Performance and Power Optimizations* Confguraton Management n Mult-Context Reconfgurable Systems for Smultaneous Performance and Power Optmzatons* Rafael Maestre, Mlagros Fernandez Departamento de Arqutectura de Computadores y Automátca Unversdad

More information

Wishing you all a Total Quality New Year!

Wishing you all a Total Quality New Year! Total Qualty Management and Sx Sgma Post Graduate Program 214-15 Sesson 4 Vnay Kumar Kalakband Assstant Professor Operatons & Systems Area 1 Wshng you all a Total Qualty New Year! Hope you acheve Sx sgma

More information

Conditional Speculative Decimal Addition*

Conditional Speculative Decimal Addition* Condtonal Speculatve Decmal Addton Alvaro Vazquez and Elsardo Antelo Dep. of Electronc and Computer Engneerng Unv. of Santago de Compostela, Span Ths work was supported n part by Xunta de Galca under grant

More information

Range images. Range image registration. Examples of sampling patterns. Range images and range surfaces

Range images. Range image registration. Examples of sampling patterns. Range images and range surfaces Range mages For many structured lght scanners, the range data forms a hghly regular pattern known as a range mage. he samplng pattern s determned by the specfc scanner. Range mage regstraton 1 Examples

More information

Comparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments

Comparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments Comparson of Heurstcs for Schedulng Independent Tasks on Heterogeneous Dstrbuted Envronments Hesam Izakan¹, Ath Abraham², Senor Member, IEEE, Václav Snášel³ ¹ Islamc Azad Unversty, Ramsar Branch, Ramsar,

More information

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana

More information

Chapter 6 Programmng the fnte element method Inow turn to the man subject of ths book: The mplementaton of the fnte element algorthm n computer programs. In order to make my dscusson as straghtforward

More information

Distributed Resource Scheduling in Grid Computing Using Fuzzy Approach

Distributed Resource Scheduling in Grid Computing Using Fuzzy Approach Dstrbuted Resource Schedulng n Grd Computng Usng Fuzzy Approach Shahram Amn, Mohammad Ahmad Computer Engneerng Department Islamc Azad Unversty branch Mahallat, Iran Islamc Azad Unversty branch khomen,

More information

A Unified Framework for Semantics and Feature Based Relevance Feedback in Image Retrieval Systems

A Unified Framework for Semantics and Feature Based Relevance Feedback in Image Retrieval Systems A Unfed Framework for Semantcs and Feature Based Relevance Feedback n Image Retreval Systems Ye Lu *, Chunhu Hu 2, Xngquan Zhu 3*, HongJang Zhang 2, Qang Yang * School of Computng Scence Smon Fraser Unversty

More information

Assembler. Shimon Schocken. Spring Elements of Computing Systems 1 Assembler (Ch. 6) Compiler. abstract interface.

Assembler. Shimon Schocken. Spring Elements of Computing Systems 1 Assembler (Ch. 6) Compiler. abstract interface. IDC Herzlya Shmon Schocken Assembler Shmon Schocken Sprng 2005 Elements of Computng Systems 1 Assembler (Ch. 6) Where we are at: Human Thought Abstract desgn Chapters 9, 12 abstract nterface H.L. Language

More information

Concurrent Apriori Data Mining Algorithms

Concurrent Apriori Data Mining Algorithms Concurrent Apror Data Mnng Algorthms Vassl Halatchev Department of Electrcal Engneerng and Computer Scence York Unversty, Toronto October 8, 2015 Outlne Why t s mportant Introducton to Assocaton Rule Mnng

More information

Algorithmic Transformation Techniques for Efficient Exploration of Alternative Application Instances

Algorithmic Transformation Techniques for Efficient Exploration of Alternative Application Instances In: Proc. 0th Int. Symposum on Hardware/Software Codesgn (CODES 02), Estes Park, Colorado, USA, May 6 8, 2002 Algorthmc Transformaton Technques for Effcent Exploraton of Alternatve Applcaton Instances

More information

An Image Fusion Approach Based on Segmentation Region

An Image Fusion Approach Based on Segmentation Region Rong Wang, L-Qun Gao, Shu Yang, Yu-Hua Cha, and Yan-Chun Lu An Image Fuson Approach Based On Segmentaton Regon An Image Fuson Approach Based on Segmentaton Regon Rong Wang, L-Qun Gao, Shu Yang 3, Yu-Hua

More information

MODULE DESIGN BASED ON INTERFACE INTEGRATION TO MAXIMIZE PRODUCT VARIETY AND MINIMIZE FAMILY COST

MODULE DESIGN BASED ON INTERFACE INTEGRATION TO MAXIMIZE PRODUCT VARIETY AND MINIMIZE FAMILY COST INTERNATIONAL CONFERENCE ON ENGINEERING DESIGN, ICED 07 28-31 AUGUST 2007, CITE DES SCIENCES ET DE L'INDUSTRIE, PARIS, FRANCE MODULE DESIGN BASED ON INTERFACE INTEGRATION TO MAIMIZE PRODUCT VARIETY AND

More information

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,

More information

User Authentication Based On Behavioral Mouse Dynamics Biometrics

User Authentication Based On Behavioral Mouse Dynamics Biometrics User Authentcaton Based On Behavoral Mouse Dynamcs Bometrcs Chee-Hyung Yoon Danel Donghyun Km Department of Computer Scence Department of Computer Scence Stanford Unversty Stanford Unversty Stanford, CA

More information

An Entropy-Based Approach to Integrated Information Needs Assessment

An Entropy-Based Approach to Integrated Information Needs Assessment Dstrbuton Statement A: Approved for publc release; dstrbuton s unlmted. An Entropy-Based Approach to ntegrated nformaton Needs Assessment June 8, 2004 Wllam J. Farrell Lockheed Martn Advanced Technology

More information

Smoothing Spline ANOVA for variable screening

Smoothing Spline ANOVA for variable screening Smoothng Splne ANOVA for varable screenng a useful tool for metamodels tranng and mult-objectve optmzaton L. Rcco, E. Rgon, A. Turco Outlne RSM Introducton Possble couplng Test case MOO MOO wth Game Theory

More information

Feature Reduction and Selection

Feature Reduction and Selection Feature Reducton and Selecton Dr. Shuang LIANG School of Software Engneerng TongJ Unversty Fall, 2012 Today s Topcs Introducton Problems of Dmensonalty Feature Reducton Statstc methods Prncpal Components

More information

Related-Mode Attacks on CTR Encryption Mode

Related-Mode Attacks on CTR Encryption Mode Internatonal Journal of Network Securty, Vol.4, No.3, PP.282 287, May 2007 282 Related-Mode Attacks on CTR Encrypton Mode Dayn Wang, Dongda Ln, and Wenlng Wu (Correspondng author: Dayn Wang) Key Laboratory

More information

Outline. Digital Systems. C.2: Gates, Truth Tables and Logic Equations. Truth Tables. Logic Gates 9/8/2011

Outline. Digital Systems. C.2: Gates, Truth Tables and Logic Equations. Truth Tables. Logic Gates 9/8/2011 9/8/2 2 Outlne Appendx C: The Bascs of Logc Desgn TDT4255 Computer Desgn Case Study: TDT4255 Communcaton Module Lecture 2 Magnus Jahre 3 4 Dgtal Systems C.2: Gates, Truth Tables and Logc Equatons All sgnals

More information

A RECONFIGURABLE ARCHITECTURE FOR MULTI-GIGABIT SPEED CONTENT-BASED ROUTING. James Moscola, Young H. Cho, John W. Lockwood

A RECONFIGURABLE ARCHITECTURE FOR MULTI-GIGABIT SPEED CONTENT-BASED ROUTING. James Moscola, Young H. Cho, John W. Lockwood A RECONFIGURABLE ARCHITECTURE FOR MULTI-GIGABIT SPEED CONTENT-BASED ROUTING James Moscola, Young H. Cho, John W. Lockwood Dept. of Computer Scence and Engneerng Washngton Unversty, St. Lous, MO {jmm5,

More information

Scheduling Remote Access to Scientific Instruments in Cyberinfrastructure for Education and Research

Scheduling Remote Access to Scientific Instruments in Cyberinfrastructure for Education and Research Schedulng Remote Access to Scentfc Instruments n Cybernfrastructure for Educaton and Research Je Yn 1, Junwe Cao 2,3,*, Yuexuan Wang 4, Lanchen Lu 1,3 and Cheng Wu 1,3 1 Natonal CIMS Engneerng and Research

More information

IMAGE FUSION TECHNIQUES

IMAGE FUSION TECHNIQUES Int. J. Chem. Sc.: 14(S3), 2016, 812-816 ISSN 0972-768X www.sadgurupublcatons.com IMAGE FUSION TECHNIQUES A Short Note P. SUBRAMANIAN *, M. SOWNDARIYA, S. SWATHI and SAINTA MONICA ECE Department, Aarupada

More information

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for

More information

Review of approximation techniques

Review of approximation techniques CHAPTER 2 Revew of appromaton technques 2. Introducton Optmzaton problems n engneerng desgn are characterzed by the followng assocated features: the objectve functon and constrants are mplct functons evaluated

More information

PHYSICS-ENHANCED L-SYSTEMS

PHYSICS-ENHANCED L-SYSTEMS PHYSICS-ENHANCED L-SYSTEMS Hansrud Noser 1, Stephan Rudolph 2, Peter Stuck 1 1 Department of Informatcs Unversty of Zurch, Wnterthurerstr. 190 CH-8057 Zurch Swtzerland noser(stuck)@f.unzh.ch, http://www.f.unzh.ch/~noser(~stuck)

More information

S1 Note. Basis functions.

S1 Note. Basis functions. S1 Note. Bass functons. Contents Types of bass functons...1 The Fourer bass...2 B-splne bass...3 Power and type I error rates wth dfferent numbers of bass functons...4 Table S1. Smulaton results of type

More information

NAG Fortran Library Chapter Introduction. G10 Smoothing in Statistics

NAG Fortran Library Chapter Introduction. G10 Smoothing in Statistics Introducton G10 NAG Fortran Lbrary Chapter Introducton G10 Smoothng n Statstcs Contents 1 Scope of the Chapter... 2 2 Background to the Problems... 2 2.1 Smoothng Methods... 2 2.2 Smoothng Splnes and Regresson

More information

PYTHON IMPLEMENTATION OF VISUAL SECRET SHARING SCHEMES

PYTHON IMPLEMENTATION OF VISUAL SECRET SHARING SCHEMES PYTHON IMPLEMENTATION OF VISUAL SECRET SHARING SCHEMES Ruxandra Olmd Faculty of Mathematcs and Computer Scence, Unversty of Bucharest Emal: ruxandra.olmd@fm.unbuc.ro Abstract Vsual secret sharng schemes

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining the Optimal Bandwidth Based on Multi-criterion Fusion Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn

More information

Active Contours/Snakes

Active Contours/Snakes Actve Contours/Snakes Erkut Erdem Acknowledgement: The sldes are adapted from the sldes prepared by K. Grauman of Unversty of Texas at Austn Fttng: Edges vs. boundares Edges useful sgnal to ndcate occludng

More information

Analysis of Continuous Beams in General

Analysis of Continuous Beams in General Analyss of Contnuous Beams n General Contnuous beams consdered here are prsmatc, rgdly connected to each beam segment and supported at varous ponts along the beam. onts are selected at ponts of support,

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Proceedngs of the Wnter Smulaton Conference M E Kuhl, N M Steger, F B Armstrong, and J A Jones, eds A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Mark W Brantley Chun-Hung

More information

Modeling, Manipulating, and Visualizing Continuous Volumetric Data: A Novel Spline-based Approach

Modeling, Manipulating, and Visualizing Continuous Volumetric Data: A Novel Spline-based Approach Modelng, Manpulatng, and Vsualzng Contnuous Volumetrc Data: A Novel Splne-based Approach Jng Hua Center for Vsual Computng, Department of Computer Scence SUNY at Stony Brook Talk Outlne Introducton and

More information

Ontology Generator from Relational Database Based on Jena

Ontology Generator from Relational Database Based on Jena Computer and Informaton Scence Vol. 3, No. 2; May 2010 Ontology Generator from Relatonal Database Based on Jena Shufeng Zhou (Correspondng author) College of Mathematcs Scence, Laocheng Unversty No.34

More information

SENSITIVITY ANALYSIS IN LINEAR PROGRAMMING USING A CALCULATOR

SENSITIVITY ANALYSIS IN LINEAR PROGRAMMING USING A CALCULATOR SENSITIVITY ANALYSIS IN LINEAR PROGRAMMING USING A CALCULATOR Judth Aronow Rchard Jarvnen Independent Consultant Dept of Math/Stat 559 Frost Wnona State Unversty Beaumont, TX 7776 Wnona, MN 55987 aronowju@hal.lamar.edu

More information

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach Angle Estmaton and Correcton of Hand Wrtten, Textual and Large areas of Non-Textual Document Images: A Novel Approach D.R.Ramesh Babu Pyush M Kumat Mahesh D Dhannawat PES Insttute of Technology Research

More information

Towards Autonomous Service Composition in A Grid Environment

Towards Autonomous Service Composition in A Grid Environment Towards Autonomous Servce Composton n A Grd Envronment Wllam K. Cheung +, Jmng Lu +, Kevn H. Tsang +, Raymond K. Wong ++ Department of Computer Scence + Hong Kong Baptst Unversty Hong Kong {wllam,jmng,hhtsang}@comp.hkbu.edu.hk

More information

Parallel Inverse Halftoning by Look-Up Table (LUT) Partitioning

Parallel Inverse Halftoning by Look-Up Table (LUT) Partitioning Parallel Inverse Halftonng by Look-Up Table (LUT) Parttonng Umar F. Sddq and Sadq M. Sat umar@ccse.kfupm.edu.sa, sadq@kfupm.edu.sa KFUPM Box: Department of Computer Engneerng, Kng Fahd Unversty of Petroleum

More information

Problem Set 3 Solutions

Problem Set 3 Solutions Introducton to Algorthms October 4, 2002 Massachusetts Insttute of Technology 6046J/18410J Professors Erk Demane and Shaf Goldwasser Handout 14 Problem Set 3 Solutons (Exercses were not to be turned n,

More information

Multiblock method for database generation in finite element programs

Multiblock method for database generation in finite element programs Proc. of the 9th WSEAS Int. Conf. on Mathematcal Methods and Computatonal Technques n Electrcal Engneerng, Arcachon, October 13-15, 2007 53 Multblock method for database generaton n fnte element programs

More information

Rapid Development of High Performance Floating-Point Pipelines for Scientific Simulation 1

Rapid Development of High Performance Floating-Point Pipelines for Scientific Simulation 1 Rapd Development of Hgh Performance Floatng-Pont Ppelnes for Scentfc Smulaton 1 G. Lenhart, A. Kugel and R. Männer Dept. for Computer Scence V, Unversty of Mannhem, B6-26B, D-68131 Mannhem, Germany {lenhart,kugel,maenner}@t.un-mannhem.de

More information

3D Virtual Eyeglass Frames Modeling from Multiple Camera Image Data Based on the GFFD Deformation Method

3D Virtual Eyeglass Frames Modeling from Multiple Camera Image Data Based on the GFFD Deformation Method NICOGRAPH Internatonal 2012, pp. 114-119 3D Vrtual Eyeglass Frames Modelng from Multple Camera Image Data Based on the GFFD Deformaton Method Norak Tamura, Somsangouane Sngthemphone and Katsuhro Ktama

More information

Composition of UML Described Refactoring Rules *

Composition of UML Described Refactoring Rules * Composton of UML Descrbed Refactorng Rules * Slavsa Markovc Swss Federal Insttute of Technology Department of Computer Scence Software Engneerng Laboratory 05 Lausanne-EPFL Swtzerland e-mal: Slavsa.Markovc@epfl.ch

More information

Hybrid Non-Blind Color Image Watermarking

Hybrid Non-Blind Color Image Watermarking Hybrd Non-Blnd Color Image Watermarkng Ms C.N.Sujatha 1, Dr. P. Satyanarayana 2 1 Assocate Professor, Dept. of ECE, SNIST, Yamnampet, Ghatkesar Hyderabad-501301, Telangana 2 Professor, Dept. of ECE, AITS,

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning the Kernel Parameters in Kernel Minimum Distance Classifier Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department

More information

Hermite Splines in Lie Groups as Products of Geodesics

Hermite Splines in Lie Groups as Products of Geodesics Hermte Splnes n Le Groups as Products of Geodescs Ethan Eade Updated May 28, 2017 1 Introducton 1.1 Goal Ths document defnes a curve n the Le group G parametrzed by tme and by structural parameters n the

More information

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 An Iteratve Soluton Approach to Process Plant Layout usng Mxed

More information

A New Token Allocation Algorithm for TCP Traffic in Diffserv Network

A New Token Allocation Algorithm for TCP Traffic in Diffserv Network A New Token Allocaton Algorthm for TCP Traffc n Dffserv Network A New Token Allocaton Algorthm for TCP Traffc n Dffserv Network S. Sudha and N. Ammasagounden Natonal Insttute of Technology, Truchrappall,

More information

Classifier Selection Based on Data Complexity Measures *

Classifier Selection Based on Data Complexity Measures * Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.

More information

Repeater Insertion for Two-Terminal Nets in Three-Dimensional Integrated Circuits

Repeater Insertion for Two-Terminal Nets in Three-Dimensional Integrated Circuits Repeater Inserton for Two-Termnal Nets n Three-Dmensonal Integrated Crcuts Hu Xu, Vasls F. Pavlds, and Govann De Mchel LSI - EPFL, CH-5, Swtzerland, {hu.xu,vasleos.pavlds,govann.demchel}@epfl.ch Abstract.

More information

Support Vector Machines

Support Vector Machines /9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.

More information

A Concurrent Non-Recursive Textured Algorithm for Distributed Multi-Utility State Estimation

A Concurrent Non-Recursive Textured Algorithm for Distributed Multi-Utility State Estimation 1 A Concurrent Non-ecursve Textured Algorthm for Dstrbuted Mult-Utlty State Estmaton Garng M. Huang, Senor Member, IEEE, and Jansheng Le, Student Member, IEEE Abstract: Durng power deregulaton, power companes

More information

Decision Strategies for Rating Objects in Knowledge-Shared Research Networks

Decision Strategies for Rating Objects in Knowledge-Shared Research Networks Decson Strateges for Ratng Objects n Knowledge-Shared Research etwors ALEXADRA GRACHAROVA *, HAS-JOACHM ER **, HASSA OUR ELD ** OM SUUROE ***, HARR ARAKSE *** * nsttute of Control and System Research,

More information

Run-Time Energy Estimation in System-On-a-Chip Designs *

Run-Time Energy Estimation in System-On-a-Chip Designs * Run-Tme Energy Estmaton n System-On-a-Chp Desgns * J. Had, G. Kaefer, Ch. Steger, R. Wess Insttute for Techncal Informatcs Graz Unversty of Technology Graz, Austra Abstract - In ths paper, a co-processor

More information

Harvard University CS 101 Fall 2005, Shimon Schocken. Assembler. Elements of Computing Systems 1 Assembler (Ch. 6)

Harvard University CS 101 Fall 2005, Shimon Schocken. Assembler. Elements of Computing Systems 1 Assembler (Ch. 6) Harvard Unversty CS 101 Fall 2005, Shmon Schocken Assembler Elements of Computng Systems 1 Assembler (Ch. 6) Why care about assemblers? Because Assemblers employ some nfty trcks Assemblers are the frst

More information

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data Malaysan Journal of Mathematcal Scences 11(S) Aprl : 35 46 (2017) Specal Issue: The 2nd Internatonal Conference and Workshop on Mathematcal Analyss (ICWOMA 2016) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES

More information

Reliability and Performance Models for Grid Computing

Reliability and Performance Models for Grid Computing Relablty and Performance Models for Grd Computng Yuan-Shun Da,2, Jack Dongarra,3,4 Department of Electrcal Engneerng and Computer Scence, Unversty of Tennessee, Knoxvlle 2 Department of Industral and Informaton

More information

Virtual Machine Migration based on Trust Measurement of Computer Node

Virtual Machine Migration based on Trust Measurement of Computer Node Appled Mechancs and Materals Onlne: 2014-04-04 ISSN: 1662-7482, Vols. 536-537, pp 678-682 do:10.4028/www.scentfc.net/amm.536-537.678 2014 Trans Tech Publcatons, Swtzerland Vrtual Machne Mgraton based on

More information

Memory Modeling in ESL-RTL Equivalence Checking

Memory Modeling in ESL-RTL Equivalence Checking 11.4 Memory Modelng n ESL-RTL Equvalence Checkng Alfred Koelbl 2025 NW Cornelus Pass Rd. Hllsboro, OR 97124 koelbl@synopsys.com Jerry R. Burch 2025 NW Cornelus Pass Rd. Hllsboro, OR 97124 burch@synopsys.com

More information

Motivation. EE 457 Unit 4. Throughput vs. Latency. Performance Depends on View Point?! Computer System Performance. An individual user wants to:

Motivation. EE 457 Unit 4. Throughput vs. Latency. Performance Depends on View Point?! Computer System Performance. An individual user wants to: 4.1 4.2 Motvaton EE 457 Unt 4 Computer System Performance An ndvdual user wants to: Mnmze sngle program executon tme A datacenter owner wants to: Maxmze number of Mnmze ( ) http://e-tellgentnternetmarketng.com/webste/frustrated-computer-user-2/

More information

Overview. Basic Setup [9] Motivation and Tasks. Modularization 2008/2/20 IMPROVED COVERAGE CONTROL USING ONLY LOCAL INFORMATION

Overview. Basic Setup [9] Motivation and Tasks. Modularization 2008/2/20 IMPROVED COVERAGE CONTROL USING ONLY LOCAL INFORMATION Overvew 2 IMPROVED COVERAGE CONTROL USING ONLY LOCAL INFORMATION Introducton Mult- Smulator MASIM Theoretcal Work and Smulaton Results Concluson Jay Wagenpfel, Adran Trachte Motvaton and Tasks Basc Setup

More information

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints Australan Journal of Basc and Appled Scences, 2(4): 1204-1208, 2008 ISSN 1991-8178 Sum of Lnear and Fractonal Multobjectve Programmng Problem under Fuzzy Rules Constrants 1 2 Sanjay Jan and Kalash Lachhwan

More information

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning Outlne Artfcal Intellgence and ts applcatons Lecture 8 Unsupervsed Learnng Professor Danel Yeung danyeung@eee.org Dr. Patrck Chan patrckchan@eee.org South Chna Unversty of Technology, Chna Introducton

More information

Vectorization of Image Outlines Using Rational Spline and Genetic Algorithm

Vectorization of Image Outlines Using Rational Spline and Genetic Algorithm 01 Internatonal Conference on Image, Vson and Computng (ICIVC 01) IPCSIT vol. 50 (01) (01) IACSIT Press, Sngapore DOI: 10.776/IPCSIT.01.V50.4 Vectorzaton of Image Outlnes Usng Ratonal Splne and Genetc

More information

Reducing Frame Rate for Object Tracking

Reducing Frame Rate for Object Tracking Reducng Frame Rate for Object Trackng Pavel Korshunov 1 and We Tsang Oo 2 1 Natonal Unversty of Sngapore, Sngapore 11977, pavelkor@comp.nus.edu.sg 2 Natonal Unversty of Sngapore, Sngapore 11977, oowt@comp.nus.edu.sg

More information

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng

More information

The stream cipher MICKEY-128 (version 1) Algorithm specification issue 1.0

The stream cipher MICKEY-128 (version 1) Algorithm specification issue 1.0 The stream cpher MICKEY-128 (verson 1 Algorthm specfcaton ssue 1. Steve Babbage Vodafone Group R&D, Newbury, UK steve.babbage@vodafone.com Matthew Dodd Independent consultant matthew@mdodd.net www.mdodd.net

More information

Cognitive Radio Resource Management Using Multi-Agent Systems

Cognitive Radio Resource Management Using Multi-Agent Systems Cogntve Rado Resource Management Usng Mult- Systems Jang Xe, Ivan Howtt, and Anta Raja Department of Electrcal and Computer Engneerng Department of Software and Informaton Systems The Unversty of North

More information