Presentation is loading. Please wait.

Presentation is loading. Please wait.

© 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 1 VLSI Architectures 048878 Lecture 3 S&F Ch. 5: Handshake Ckt Implementations.

Similar presentations


Presentation on theme: "© 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 1 VLSI Architectures 048878 Lecture 3 S&F Ch. 5: Handshake Ckt Implementations."— Presentation transcript:

1 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 1 VLSI Architectures 048878 Lecture 3 S&F Ch. 5: Handshake Ckt Implementations

2 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 2 Implementations We only consider simple circuits More aggressive circuits will come later First, reminder on latches

3 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 3 4- & 2-phase bundled data latches

4 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 4 4-phase dual rail – many bits

5 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 5 4-phase Fork, Join

6 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 6 4-phase Bundled-data Mux

7 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 7 4-phase Bundled-data Demux

8 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 8 4-phase Merge

9 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 9 4-phase Merge Mutually exclusive inputs. Guaranteed elsewhere! (more later..) Assume X active… …C-element sees input glitch Relative Timing: x-req  < z-ack   simplify CEL

10 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 10 Asymmetric C Element Useful when we know the relative timing: b  < a   only a  needed to pull up Only one pMOS - faster

11 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 11 2-phase Merge Try it at home… This is not an assignment!

12 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 12 Mutual Exclusion: MUTEX

13 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 13 Standard Gate MUTEXs Not fully guaranteed that outputs are M/E, but highly probable ! Very low threshold

14 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 14 Arbiter

15 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 15 Arbitrating Merge

16 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 16 Function Blocks We said “transparent” but… –Need a matched delay for bundled-data –Need to generate completion for dual-rail –Need to join inputs, fork outputs:

17 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 17 Transparency Revisited Function blocks must not affect how the latches “shake hands” (except for timing)

18 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 18 Indication Revisited FB(req_out  ) means –FB(req_in  ) –Computation finished, data out ready Simple “strong indication” for bundled data: 1: ALL DATA_IN VALID 2: REQ_IN  3: COMPUTE 4: ALL DATA_OUT VALID 5: REQ_OUT 

19 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 19 Strong vs. Weak Indication Strong Indication: All inputs must arrive before any output is allowed (“indicated”). –Even if some outputs are ready earlier, there is no REQ_OUT, so they cannot be used. –Implies worst-case latency Weak Indication: Some outputs are allowed even before all inputs arrived –Only makes sense in dual-rail:

20 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 20 Weak Indication No REQ on dual-rail – each bit is “self- indicating” May lead to faster circuits Example chain of events: 1 2 34 6 5 7

21 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 21 Composition of FBs Legal composition: –All inputs and outputs are connected –No cycles Legal composition of weekly indicating FBs is weakly indicating Legal composition of strongly indicating FBs is strongly indicating

22 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 22 Example: Ripple-carry

23 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 23 Example: Ripple-carry Full adder (a,b,c) = (s,d) –s = a  b  c –d = ab + ac + bc Shortcuts for look-ahead (prop, gen, kill): –p = a  b s = p  c –g = abd = g + pc, OR d' = k + pc' –k = a' b' Sometimes d can be made valid without waiting for c

24 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 24 Speculative / Strong Ripple Carry 16 bit ripple-carry adder, bundled-data Longest carry is 16 stages But if p 8 =0 then longest carry is 8 stages And if p 12 p 8 p 4 =0, then longest carry is 4 stages If willing to trade area and power for speed:

25 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 25 Speculative / Strong Ripple Carry

26 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 26 ST-CL Based on David, Ginosar, Yoeli, "An Efficient Implementation of Boolean Functions as Self-Timed Circuits,'' IEEE Trans. Computers, Jan. 1992

27 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 27 Dual-Rail DIMS PLA Notation

28 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 28 Dual-Rail DIMS Adders Still slow: LF(V) = LF(E)

29 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 29 Transistor Level DIMS Too many P transistors - slow Some N paths can be shared:

30 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 30 Hybrid Adder Dual-rail carry (for flexible latency) Bundled-data data inputs and sum output (for lower area and power) Data-dependent data-forward (V) latency Constant empty-forward (E) latency

31 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 31 Hybrid Adder Dual-rail Bundled-data

32 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 32 Domino Logic Dual Rail Req Out: Either by (flexible) Completion Detection or by matched (worst case) delay

33 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 33 Hybrid Adder: Sum Ckt

34 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 34 Hybrid Adder: Two Carry Ckts Weak IndicationStrong Indication KILL GEN

35 © 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 35 Hybrid Adder: Two Carry Ckts WEAK CARRY STRONG CARRY STRONG CARRY WEAK CARRY STRONG CARRY STRONG CARRY … 123456 CD Slightly faster…


Download ppt "© 2003-2009 Ran Ginosar048878 Lecture 3: Handshake Ckt Implementations 1 VLSI Architectures 048878 Lecture 3 S&F Ch. 5: Handshake Ckt Implementations."

Similar presentations


Ads by Google