Skip to content
DenseChip
Learn
PracticeCode labInterview

DenseChip

Navigate

Menu

LearnPracticeCode labInterview

Tools

All toolsSV editorCode repoResume
DenseChip

Practice, courses, and chip-design tools in one focused workspace.

LearnPracticeCode labInterviewToolsLearning paths

© 2026 DenseChip

AboutContactPrivacyTerms
Skip to lesson

Course contents

  • Logic Synthesis Glossary
  • Logic Synthesis Cheatsheet
  • Synthesis Foundations14
  • Logic Optimisation and Mapping20
  • Constraints and Timing15
    • SDC for Synthesis
    • Delay Models and Slack
    • Timing-Driven Synthesis
      • Critical-Path Focus
      • Restructuring and Duplication
      • Sizing and Path Groups
      • Reading WNS and TNS
      • Margin and When to Stop
  • Sequential, FSM and Memory19
  • Datapath Synthesis14
  • Area, Power and Test14
  • Verification, ECO and Handoff15
  • Flows and Debugging15
  • Interview and Judgment11
Loading lesson

‹Timing-Driven Synthesis
  1. Topics
  2. Logic Synthesis
  3. Constraints and Timing
  4. Timing-Driven Synthesis
  5. Critical-Path Focus

Part 3 · Constraints and Timing · Intermediate

Critical-Path Focus

Slack ranks paths so synthesis spends effort on the critical path first.

On this page

    Slack ranks every path

    Slack is required time minus arrival time on one path. The path with the smallest (worst) slack is the critical path.

    When all slacks are positive, no path violates, but a critical path still exists.

    A synthesis tool recomputes these ranks as the netlist changes. Each round of effort targets the current worst path.

    Diagram
    EXAMPLE parameters: required time 6.0 delay units
    path arrival = 4.0 + 3.0 + 3.0 = 10.0 delay units
    slack = required - arrival = 6.0 - 10.0 = -4.0 delay units
    negative slack means the path misses its target by 4.0 delay units

    Effort follows the worst path

    Timing-driven synthesis applies its transforms to the critical path first. Restructuring, sizing, and duplication each aim at the same ranked list.

    Fixing a non-critical path changes no WNS; it can still improve TNS. Its value shows in TNS until that path becomes the worst one.

    The timing-driven loop.
    Each pass re-ranks the paths and spends the next effort on the new worst one.

    Work in progress versus final quality

    Intermediate runs explore structures, so their slack can move either way between runs. A worse middle run does not prove the approach is wrong.

    Final quality is read at the real target after effort converges. Judge a strategy by its trend across runs, not by one middle number.

    Common mistakes

    Common pitfall

    Polishing one path far past zero slack while others still fail wastes area. Move effort to the new worst path once the current one meets its target.

    Interview

    Likely probes: what the critical path is and where effort goes.

    • The critical path is the path with the smallest slack, even when every slack is positive.

    • Effort follows the ranked list; non-critical fixes leave WNS but can improve TNS.

    • Middle runs are exploration; final quality is read at the real target.

    Check yourself Path A has slack -4.0 and path B has slack -0.5 delay units. Which path gets the next effort, and why?

    Path A. It is the critical path, and only the worst slack sets the reported worst total.

    Check yourself A middle run is worse than the run before it. Does that prove the strategy failed?

    No. Intermediate structures are exploration. The trend across runs and the final result at the target decide.

    Related topics

    • Up: Timing-Driven Synthesis

    • Next: Restructuring and Duplication

    • Delay Models and Slack

    Previous

    47 / 139

    Next: Restructuring and Duplication

    Up next

    Up nextRestructuring and Duplication

    Sign in to save progress

    Practice this lessonReport an error

    See also

    • Sizing and Path Groups
    • Reading WNS and TNS
    • Margin and When to Stop