DOC PREVIEW
Berkeley ELENG C249A - Constraint-Driven Communication Synthesis

This preview shows page 1-2-14-15-29-30 out of 30 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 30 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 30 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 30 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 30 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 30 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 30 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 30 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation1Constraint-Driven Communication SynthesisAlessandro PintoGuoqiang WangMentors:Luca Carloni12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation2Outline• Introduction• Problem Definition• Solving the Constrained Optimization Problem• Algorithm and examples• Conclusion and Future Work12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation3IntroductionA General System View:12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation4CCG12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation5CCG: • A Comm. Constraint Graph, is a directed graph, an abstract model used to specify comm. requirements.• A vertex : associated to a port of computational module of the system.• A directed arc : representing a point-to-point comm. channel btw two modules.• Arc Properties of : arc length/distance btw and .: comm. bandwidth on the constrained arc .),( AVgG=),( vua=va)(ad)(abvua12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation6A Comm. Library• A collection of comm. links and comm. nodes. • Link Properties:: link length/distance: the length of the longest comm. channel that can be realized by this link.: link bandwidth: the bandwidth of the fastest communication channel that can be realized by this link.: link cost: the value of this link wrt to other links in the library based on an optimality criterion varying with the type of application.NLL∪=')(ld)(lc)(lb12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation7Implementation Graph• The realization of a comm. architecture satisfying the requirements specified by a constraint graph• Ports of the in the Constraints Graph are kept• For each arc in the constraint graph we have a path in the implementation graphü There is a set of new nodes in the implementation graph N’ that are instances of library elementsü Each arc in the implementation graph is an instance of a library element12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation8A path• An alternating sequence of distinct vertices and arcs in a constraint graph , with and denoting respectively the set of vertices and arcs touched by . • path properties: length: bandwidth:cost:),,,...,,,,(112211 QQQvavavavq−−=Gq),( GqV ),( GqA∑−==10)()(Qiiadqd∑−==10)()(Qiiacqc))((min)()1,...0( iQiabqb−==12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation9Arc Implementation • Arc Matching: iff corresponding to exactly one library link.• K-way Arc Segmentation: iffcorresponding to the concatenation of K library links.• K-way Arc Duplication: iff corresponding to placing K library links in parallel.12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation10Examples of AI12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation11Lemma• For all Constraint graphs and all communication libraries , there exists an optimum point-to-point implementation graph with the minimum cost.),( AVgG=NLL∪='12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation12Constraint-driven comm. synthesis problemTo find the comm. architecture satisfying all the constraints specified as comm. requirements on the channels, while minimizing a predefined cost function capturing an optimality criterion which must be defined for the specific application.12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation13Our Problem• Given: A Constraint Graph: A Comm. Library: • Find:The Optimum implementation graph with the minimum cost),( AVgG=NLL∪='12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation14Assumption on L• Links that have better performance cost more)'()()'()()'()(',lclclblbldldLll≤⇒≤∧≤∈12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation15Idea• How can we solve the problem?üGenerate all the solutions?• And then?üAsk to the mentors?• And then? Got the solution• I’m a Synthesis expert, if you give me all the possible solutions I can solve the problem as a UCP12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation16Complex Solution• Consider all the possible structuresGG’G’’G’’’G’’’’G’’’’’12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation17K-way mergeability• K arcs are mergeable is k-way merge is the minimumcost solution12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation18Non-mergeability Conditions• Bandwidth limitation üIt has to be proved• Conditions on relative positions of arcs12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation19Simple case: 2-way MergingmergeablenovpvpupupadadifAaa−⇒−+−≤+∈||)'()(||||)'()(||)'()(',vsaa’a a’1) 2)12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation20K-way Merge• Same Concept, express every properties w.r.t. one link{}kkinkninininknnikkMAvpvpupupadadkAaaaA∉⇒−+−≤+−⊆=∑∑≠∈≠∈ ],1[],1[21||)()(||||)()(||)()()1(...,,12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation21Bandwidth Condition (intuition)Ga1a2a3akG’We need Duplication to satisfy the bandwidth constraintG’r-way merge plus (k-r)-way merge12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation22Non-Mergeability Theorem for Bandwidth • Duplication is allowed only if there is no link in the common subpath the doesn’t share the implementation of some arcs {}{ }{ }kkkijikjLlikkMAvublbabaaA∉⇒+≥=∑=∈∈1],1[1),(min)(max)(,...,12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation23Non-Mergeability Theorem for Relative Positions (intuition)• Suppose that a cannot be merged with no other arcs in AüWhat does it means?üWhat can we say about merging a with 2 arcs?2222,, MAAaA ∉∈∀?333MAaA∈⊃12/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation24Non-Mergeability Theorem for Relative Positions• If a is not k-way mergeable with any set of k ark in A it is not k+1 mergeable with any set of k+1-arc in A{}{}(){ } { }( )hhhkkkMaAaAAAkhMaAaAA∉∪⊂∀∈∀⇒∉∪⊂∀−−,\|],|,[,\1112/8/2001Alessandro PintoGuoqiang WangEE249 Final Project Presentation25What we did so far• Mergeability definition• Non-Mergeability ConditionsüRelative


View Full Document

Berkeley ELENG C249A - Constraint-Driven Communication Synthesis

Documents in this Course
Load more
Download Constraint-Driven Communication Synthesis
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Constraint-Driven Communication Synthesis and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Constraint-Driven Communication Synthesis 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?