jagomart
digital resources
picture1_Work Spreadsheet 42315 | C304 Slide


 242x       Filetype PPTX       File size 0.84 MB       Source: vlsicad.ucsd.edu


File: Work Spreadsheet 42315 | C304 Slide
outline gate sizing in vlsi design previous work challenges in gate sizing high performance gate sizing with a signoff timer overall flow experimental results conclusions and future work 2 gate ...

icon picture PPTX Filetype Power Point PPTX | Posted on 16 Aug 2022 | 3 years ago
Partial capture of text on file.
  Outline
  • Gate Sizing in VLSI Design
  • Previous Work
  • Challenges in Gate Sizing
  • High-Performance Gate Sizing with a Signoff Timer
  • Overall Flow
  • Experimental Results
  • Conclusions and Future Work
                                                              2
   Gate Sizing in VLSI Design
   • Effective approach to power, delay optimization 
   • Objective: minimize  power 
     • Satisfy constraints: slack, slew, max load capacitance, …
     • Tunable cell parameters: gate width, Vth, gate length
     • Select a proper library cell for each gate
             gate-width
             gate-width          INVX2     INVX4    INVX8    INVX16
           (drive-strength)
           (drive-strength)
             multi-Vth           HVT            NVT           LVT
             multi-Vth
              L  -bias
              Lgate-bias        L=65nm         L=60nm       L=55nm
               gate
                  lower (leakage) power            higher (leakage) power
                  lower speed                      higher speed
                                                                           3
   Previous Gate Sizing Techniques
   • Common heuristics/algorithms
                                                    Convex optimization
       Continuous      Linear  programming          Convex optimization
       Continuous      Linear  programming
       gate sizing
       gate sizing                                 Lagrangian relaxation
                                                   Lagrangian relaxation
        Discrete 
        Discrete      Dynamic programming          Sensitivity-based sizing
       gate sizing    Dynamic programming          Sensitivity-based sizing
       gate sizing
   • Limitations
     • Continuous gate sizing : industrial cell libraries have discrete gate 
        sizes, and rounding solutions may be suboptimal
     • Discrete gate sizing : NP-hard problem  scalability issue
     • Do not account for realistic delay models and constraints (capaci-
        tance, slew)
                                                                             4
  Previous Work
  • Our work extends Trident 1.0 
    [Hu et al. Proc. ICCAD 2012]
    • Produced strongest results on ISPD 2012 benchmarks
      as of ICCAD 2012
    • Metaheuristic optimization with importance sampling 
      and sensitivity-guided search
    • Limitation: no interconnect delay calculation
      ⇒ Unrealistic assumption
                                                             5
  Outline
  • Gate Sizing in VLSI Design
  • Previous Work
  • Challenges in Gate Sizing
    • Issue 1: Interconnect delay
    • Issue 2: Inaccurate internal timer
    • Issue 3: Critical paths
  • High-Performance Gate Sizing with a Signoff Timer
  • Overall Flow
  • Experimental Results
  • Conclusions and Future Work
                                                              6
The words contained in this file might help you see if this file matches what you are looking for:

...Outline gate sizing in vlsi design previous work challenges high performance with a signoff timer overall flow experimental results conclusions and future effective approach to power delay optimization objective minimize satisfy constraints slack slew max load capacitance tunable cell parameters width vth length select proper library for each invx drive strength multi hvt nvt lvt l bias lgate nm lower leakage higher speed techniques common heuristics algorithms convex continuous linear programming lagrangian relaxation discrete dynamic sensitivity based limitations industrial libraries have sizes rounding solutions may be suboptimal np hard problem scalability issue do not account realistic models capaci tance our extends trident produced strongest on ispd benchmarks as of iccad metaheuristic importance sampling guided search limitation no interconnect calculation unrealistic assumption inaccurate internal critical paths...

no reviews yet
Please Login to review.