Communication-Based Design

Transcription

1 Communication-Based Design

2 Motivation System-level verification of large component-oriented designs will be very costly. We cannot afford to debug interface mismatches between internal components... especially considering that there will be many, many interfaces between so many components. Current situation is unacceptable Interfaces are not specified precisely No clear specification formalism exist. Tools to create, debug, and make maximal use of the specifications don t exist.

3 Basic Goals Identify precisely and formally the concept of communication, its level of abstraction and of the corresponding models of computation. New theories to combine different models of computation Determine a set of properties that characterizes each level of abstraction. Provide methods and tools to extract formal properties and specifications from informal ones and existing, ambiguously specified standards.

4 Basic Goals Formal synthesis of protocols Bus architectures analysis and verification formal specs monitors semi-formal analysis Abstract the notion of communication architectures and include estimation processes to quickly evaluate different architecture

5 Why separating computation and communication? Verification (debugging). If not: Communication hard-wired with computation Often hard to tell who is at fault Bugs may be distributed, difficult to track down Changes in the system may require rewriting of entire blocks, often leading to new bugs Reuse Component may be plugged in different environments Functions and interface behavior are difficult to separate Architecture exploration Design components with abstract communication primitives Explore different implementations without touching the component

6 What is Communication? Os = Fs(is) Or = Fr(ir) Is os ir or S R

7 What is Communication? Os = Fs(is) Connection Or = Fr(ir) Is os ir or S C R Connection C enables the interaction between the behaviors S and R

8 What is Communication? Os = Fs(is) Or = Fr (ir ) Is os Ir Or S R Ideal Connection Fs Fc Fr

9 What is Communication? Os = Fs(is) Or = Fr (ir ) Is os Ir Or S R Ideal Connection Fs Fc Fr S restricts the behavior of R to R

10 Behavior Adaptation Os = Fs(is) Or = Fr(ir) Is os ir or S Ideal Connection R Fs Fc Fr R not defined for some output of S: behavior mismatch

11 Behavior Adaptation Os = Fs(is) Or = Fr(ir) Is os Z = Z o Z ir or S R Fs Fc Fz Fr Behavior Adapter Z maps outputs of S into the domain of R

12 Behavior Adaptation Os = Fs (is ) = Fs o Z (is ) Or = Fr (ir ) = Fr o Z (ir ) Z Z Is Os Ir Or S R Behavior Adapter encapsulates S and R S and R communicate successfully over an ideal connection

13 Physical Channels Os = Fs (is ) Or = Fr (ir ) Is Os Ir Or S C R Fs Fc = id Fr Invalid Channels may introduce mismatch due to their physical properties (noise, interference ) Valid Channels satisfy QoS requirements QoS-equivalent to the ideal connection ( Fs o Fc o Fr ~ Fs o Fr )

14 Channel Adapter Os = Fs (is ) C Or = Fr (ir ) Is S SC C CR R Or Choose SC and CR such that C is valid Fs o Fsc o Fc o Fcr o Fr ~ Fs o Fr Channel Adapter may introduce behavior mismatch need a Behavior Adapter

15 FIFOs as Behavior Adapters S S R FIFOs adapt the rates of S and R Unbounded FIFOs ideal adapter Bounded FIFOs to prevent overflow, restrict S using blocking write (Req/Ack)

16 Protocol Design S R (Ideal) Connection S Z R Behavior Adaptation C = id S Z Z R Physical Channel selection C = id S Z SC CR Z R Channel Adaptation C = id S Z SC Z Z CR Z R Behavior Adaptation

17 Protocol Design S R (Ideal) Connection S C = id S R S R S Behavior Adaptation Physical Channel selection S C = id R S Channel Adaptation S C = id R Behavior Adaptation

18 Design Methodology Functional Decomposition Behavior Adaptation Communication Media Insertion MoC Wrapping Communication Refinement Channel Adaptation Mapping and Optimizations

19 Functional Decomposition Functional Decomposition at the highest abstraction level, a system is a single process System it is refined into a set of concurrent processes Process: relation between an input domain and an output codomain only behavior, no communication P3 P1 P4 P2 P5 denotational specification

20 Functional Decomposition (ex.) MPEG Decoder VLD IDCT MC DISPLAY

21 Behavior Adaptation Behavior adapters match different domains, so that processes can understand each other relation between two domains not part of original system specification: needed because of the particular decomposition needed independently of how the communication is performed P1 relation P3 P1 Input domain P4 P2 behavior adapter P2 relation P5

22 Behavior Adaptation (ex.) vect blkvect VLD IDCT MC DISPLAY block block Mblock Mblock Frame mot. vect mot. vect IZ,IQ MEM MEM VLD IDCT MC DISPLAY

23 Communication and MoC Communication medium each link needs a communication medium does not affect or change the relation inside processes P1 P3 P4 P2 P5 MoC wrapper used to establish a firing rule and a communication semantics for each process only the Moc wrapper is modified if a medium is changed P W P

24 Communication and Moc (ex.) IZ,IQ MEM MEM VLD IDCT MC DISPLAY VLD IDCT MC DISPLAY IZ,IQ MEM MEM

25 Refinement Refinement any communication medium can be refined into an arbitrary netlist, as long as the interface is not changed Channel adapters P1 P3 P4 P2 Lossless P5 used to preserve properties of a given interface example: lossless communication realized with a lossy medium ( retransmission + acknowledge) CA Lossy CA

26 Refinement (ex.) VLD IDCT MC DISPLAY IZ,IQ MEM MEM VLD IDCT MC DISPLAY IZ,IQ MEM MEM SEG SEG REAS SEG REAS REAS ARB ARB ADDR ARB ADDR ADDR BUS

27 Mapping and Optimization P1 P2 Optimization P4 map each element (processes, adapters, media) onto architecture P3 CA Lossy CA Lossless P5 merge processes, adapters and media into a single process, when applicable provide an imperative description for each process

28 Mapping and Optimization (ex.) VLD IDCT MC DISPLAY IZ,IQ SEG SEG REAS SEG REAS REAS ARB ARB ADDR ARB ADDR ADDR BUS MEM MC MEM