Measurement of Inherent Noise in EDA Tools

1
Measurement of Inherent Noise in EDA Tools

• Andrew B. Kahng
• and
• Stefanus Mantik
• UCSD CSE and ECE Departments, La Jolla, CA
• UCLA CS Department, Los Angeles, CA

2
Introduction

• Complexity ?, design cycle time ?
• Tool predictability
• predict final solution before running the tool
• requires understanding of tool behavior
• Heuristic nature of tool
• leads to noise variability in solution quality
• Understand how tool behaves in the presence of
  noise

3
Previous Works

• Hartoog (DAC86)
• reorder cells/nets ? isomorphic circuits
• algorithm comparison
• Harlow/Brglez (ICCAD98)
• randomize name and node order
• Ghosh (Dissertation00)
• circuit mutation ? isomorphic class
• Bodapati/Najm (SLIP00)
• noise effect on pre-layout estimation

4
Outline

• Tool behavior criteria
• Taxonomy of potential perturbation
• Examples on noise effects
• Exploitation of tool noise
• Conclusions and ongoing works

5
Tool Behavior Criteria

• Monotonicity
• monotone solutions w.r.t. inputs

6
Tool Behavior Criteria

• Monotonicity
• Smoothness
• similar solutions after ? perturbation

Solution space
7
Tool Behavior Criteria

• Monotonicity
• Smoothness
• Scaling
• preserve quality with scaled input

0.18?
0.25?
8
Outline

• Tool behavior criteria
• Taxonomy of potential perturbation
• Examples on noise effects
• Exploitation of tool noise
• Conclusions and ongoing works

9
Perturbation Taxonomy

• Randomness
• Ordering and naming
• Coarseness and richness of library
• Constraints
• Geometric properties

10
Randomness

• Random number generator (RNG)
• initial solution for heuristic
• tie breaker

11
Ordering and Naming

• Instance ordering
• C1,C2,C3, ? C17,C224,C5,
• Instance naming
• AFDXCTRLAX239 ? CELL00134
• AFDXCTRLAX239 ? ID012ID119ID416

12
Library Coarseness Richness

• Cell library
• number of variations for cell types (e.g., INV1x,
  INV2x, INV4x, INV8x, etc.)
• Timing library
• timing model (look-up tables, linear
  interpolation, etc.)

13
Constraints

• Design rules
• spacing, width, size, etc.
• Design constraints
• timing constraints
• grouping constraints
• area constraints
• Perturbation
• tightening or relaxing the constraints

14
Geometric Properties

• Offsets
• cell sites, cell rows, routing tracks, power
  stripes, global cell grids, etc.
• Orientations
• pin orientations, site orientations, routing
  directions, etc.
• Instance scaling
• cell sizes, routing pitches, layout size, etc.
• Artificial blockages

15
Outline

• Tool behavior criteria
• Taxonomy of potential perturbation
• Examples on noise effects
• Exploitation of tool noise
• Conclusions and ongoing works

16
Examples of Noise Effects

• Monotonicity test
• Random seeds
• Random ordering and naming
• Random hierarchy
• Cadence Place Route
• 13 industry designs

17
Monotonicity Test

• OptimizationLevel 1(fast/worst) 10(slow/best)

18
Random Seeds

• 200 runs with different random seeds
• 0.05 improvement

-0.05
19
Random Ordering Naming

• Data sorting ? no effect on reordering
• Five naming perturbation
• random cell names without hierarchy (CR)
• E.g., AFDXCTRLAX239 ? CELL00134
• random net names without hierarchy (NR)
• random cell names with hierarchy (CH)
• E.g., AFDXCTRLAX129 ? ID012ID79ID216
• random net names with hierarchy (NH)
• random master cell names (MC)
• E.g., NAND3X4 ? MCELL0123

20
Random Naming (contd.)

• Wide range of variations (3)
• Hierarchy matters

Number of Runs
Quality Difference
21
Random Hierarchy

• Swap hierarchy
• AABBC03 ? XXYYC03
• XXYYZ12 ? AABBZ12

Number of Runs
Quality Difference
22
Outline

• Tool behavior criteria
• Taxonomy of potential perturbation
• Examples on noise effects
• Exploitation of tool noise
• Conclusions and ongoing works

23
Noise Additive Property
?

• Noise1 Noise2 (Noise1 Noise2)

24
Noise Exploitation

• CPU Budget 1 run
• noise with best mean
• CPU Budget 5 runs
• noise with min average soln. over 5 runs
• For each noise
• randomly select k solutions
• record the best-k
• repeat 1000 times and get average

25
Noise Exploitation (Contd.)

• Noise that preserves hierarchy almost always
  yields superior results
• CPU budget 1 run ? use MC

26
Outline

• Tool behavior criteria
• Taxonomy of potential perturbation
• Examples on noise effects
• Exploitation of tool noise
• Conclusions and ongoing works

27
Conclusions

• EDA tools behavior criteria w.r.t. noise
• Initial taxonomy of noise sources
• Effects of noises on PR solutions
• Non-additive noise property

28
Ongoing Works

• Prediction model that includes noises
• Relationships between different noises
• Noise impact on timing-driven solution
• Relationship between perturbation size and
  changes in solution quality
• Composition of noises between consecutive tools
  in the design flow