Eyecharts: Constructive Benchmarking of Gate Sizing Heuristics Puneet Gupta, University of California, Los Angeles Andrew B. Kahng, University of California,

Slides:



Advertisements
Similar presentations
Gate Sizing for Cell Library Based Designs Shiyan Hu*, Mahesh Ketkar**, Jiang Hu* *Dept of ECE, Texas A&M University **Intel Corporation.
Advertisements

Gregory Shklover, Ben Emanuel Intel Corporation MATAM, Haifa 31015, Israel Simultaneous Clock and Data Gate Sizing Algorithm with Common Global Objective.
Design Rule Generation for Interconnect Matching Andrew B. Kahng and Rasit Onur Topaloglu {abk | rtopalog University of California, San Diego.
ELEN 468 Lecture 261 ELEN 468 Advanced Logic Design Lecture 26 Interconnect Timing Optimization.
Timing Margin Recovery With Flexible Flip-Flop Timing Model
Improving Placement under the Constant Delay Model Kolja Sulimma 1, Ingmar Neumann 1, Lukas Van Ginneken 2, Wolfgang Kunz 1 1 EE and IT Department University.
UC San Diego / VLSI CAD Laboratory NOLO: A No-Loop, Predictive Useful Skew Methodology for Improved Timing in IC Implementation Tuck-Boon Chan, Andrew.
The Cost of Fixing Hold Time Violations in Sub-threshold Circuits Yanqing Zhang, Benton Calhoun University of Virginia Motivation and Background Power.
DARPA Assessing Parameter and Model Sensitivities of Cycle-Time Predictions Using GTX u Abstract The GTX (GSRC Technology Extrapolation) system serves.
High-Performance Gate Sizing with a Signoff Timer
Minimum-Buffered Routing of Non- Critical Nets for Slew Rate and Reliability Control Supported by Cadence Design Systems, Inc. and the MARCO Gigascale.
NuCAD ELECTRICAL ENGINEERING AND COMPUTER SCIENCE McCormick Northwestern University Robert R. McCormick School of Engineering and Applied Science FA-STAC.
Layer Assignment Algorithm for RLC Crosstalk Minimization Bin Liu, Yici Cai, Qiang Zhou, Xianlong Hong Tsinghua University.
Background: Scan-Based Delay Fault Testing Sequentially apply initialization, launch test vector pairs that differ by 1-bit shift A vector pair induces.
Puneet Sharma and Puneet Gupta Prof. Andrew B. Kahng Prof. Dennis Sylvester System-Level Living Roadmap Annual Review, Sept Basic Ideas Gate-length.
Power-Aware Placement
UC San Diego Computer Engineering VLSI CAD Laboratory UC San Diego Computer Engineering VLSI CAD Laboratory UC San Diego Computer Engineering VLSI CAD.
Input-Specific Dynamic Power Optimization for VLSI Circuits Fei Hu Intel Corp. Folsom, CA 95630, USA Vishwani D. Agrawal Department of ECE Auburn University,
ABSTRACT We consider the problem of buffering a given tree with the minimum number of buffers under load cap and buffer skew constraints. Our contributions.
Architectural-Level Prediction of Interconnect Wirelength and Fanout Kwangok Jeong, Andrew B. Kahng and Kambiz Samadi UCSD VLSI CAD Laboratory
© Digital Integrated Circuits 2nd Inverter CMOS Inverter: Digital Workhorse  Best Figures of Merit in CMOS Family  Noise Immunity  Performance  Power/Buffer.
Constructing Current-Based Gate Models Based on Existing Timing Library Andrew Kahng, Bao Liu, Xu Xu UC San Diego
April 16th, Photomask Japan 2008 Electrical Metrics for Lithographic Line-End Tapering Puneet Gupta 3,
Chung-Kuan Cheng†, Andrew B. Kahng†‡,
On-Line Adjustable Buffering for Runtime Power Reduction Andrew B. Kahng Ψ Sherief Reda † Puneet Sharma Ψ Ψ University of California, San Diego † Brown.
1 UCSD VLSI CAD Laboratory ISQED-2009 Revisiting the Linear Programming Framework for Leakage Power vs. Performance Optimization Kwangok Jeong, Andrew.
Toward Performance-Driven Reduction of the Cost of RET-Based Lithography Control Dennis Sylvester Jie Yang (Univ. of Michigan,
A Cost-Driven Lithographic Correction Methodology Based on Off-the-Shelf Sizing Tools.
Detailed Placement for Leakage Reduction Using Systematic Through-Pitch Variation Andrew B. Kahng †‡ Swamy Muddu ‡ Puneet Sharma ‡ CSE † and ECE ‡ Departments,
ELEN 468 Lecture 271 ELEN 468 Advanced Logic Design Lecture 27 Interconnect Timing Optimization II.
SLIP 2000April 9, Wiring Layer Assignments with Consistent Stage Delays Andrew B. Kahng (UCLA) Dirk Stroobandt (Ghent University) Supported.
Selective Gate-Length Biasing for Cost-Effective Runtime Leakage Control Puneet Gupta 1 Andrew B. Kahng 1 Puneet Sharma 1 Dennis Sylvester 2 1 ECE Department,
Optimality Study of Logic Synthesis for LUT-Based FPGAs Jason Cong and Kirill Minkovich VLSI CAD Lab Computer Science Department University of California,
1 A Novel Metric for Interconnect Architecture Performance Parthasarathi Dasgupta, Andrew B. Kahng, Swamy V. Muddu Dept. of CSE and ECE University of California,
1 A Method for Fast Delay/Area Estimation EE219b Semester Project Mike Sheets May 16, 2000.
UC San Diego Computer Engineering VLSI CAD Laboratory UC San Diego Computer Engineering VLSI CAD Laboratory UC San Diego Computer Engineering VLSI CAD.
DDRO: A Novel Performance Monitoring Methodology Based on Design-Dependent Ring Oscillators Tuck-Boon Chan †, Puneet Gupta §, Andrew B. Kahng †‡ and Liangzhen.
Enhanced Metamodeling Techniques for High-Dimensional IC Design Estimation Problems Andrew B. Kahng, Bill Lin and Siddhartha Nath VLSI CAD LABORATORY,
UC San Diego / VLSI CAD Laboratory Reliability-Constrained Die Stacking Order in 3DICs Under Manufacturing Variability Tuck-Boon Chan, Andrew B. Kahng,
-1- UC San Diego / VLSI CAD Laboratory Methodology for Electromigration Signoff in the Presence of Adaptive Voltage Scaling Wei-Ting Jonas Chan, Andrew.
Dose Map and Placement Co-Optimization for Timing Yield Enhancement and Leakage Power Reduction Kwangok Jeong, Andrew B. Kahng, Chul-Hong Park, Hailong.
DELAY INSERTION METHOD IN CLOCK SKEW SCHEDULING BARIS TASKIN and IVAN S. KOURTEV ISPD 2005 High Performance Integrated Circuit Design Lab. Department of.
Power Reduction for FPGA using Multiple Vdd/Vth
UC San Diego / VLSI CAD Laboratory Toward Quantifying the IC Design Value of Interconnect Technology Improvement Tuck-Boon Chan, Andrew B. Kahng, Jiajia.
Horizontal Benchmark Extension for Improved Assessment of Physical CAD Research Andrew B. Kahng, Hyein Lee and Jiajia Li UC San Diego VLSI CAD Laboratory.
Logic Synthesis For Low Power CMOS Digital Design.
1 Design Space Exploration for Power-Efficient Mixed-Radix Ling Adders Chung-Kuan Cheng Computer Science and Engineering Depart. University of California,
UC San Diego / VLSI CAD Laboratory Incremental Multiple-Scan Chain Ordering for ECO Flip-Flop Insertion Andrew B. Kahng, Ilgweon Kang and Siddhartha Nath.
An Efficient Clustering Algorithm For Low Power Clock Tree Synthesis Rupesh S. Shelar Enterprise Microprocessor Group Intel Corporation, Hillsboro, OR.
-1- UC San Diego / VLSI CAD Laboratory Construction of Realistic Gate Sizing Benchmarks With Known Optimal Solutions Andrew B. Kahng, Seokhyeong Kang VLSI.
Kwangsoo Han, Andrew B. Kahng, Hyein Lee and Lutong Wang
Kwangsoo Han‡, Andrew B. Kahng‡† and Hyein Lee‡
Tao Lin Chris Chu TPL-Aware Displacement- driven Detailed Placement Refinement with Coloring Constraints ISPD ‘15.
Fast Algorithms for Slew Constrained Minimum Cost Buffering S. Hu*, C. Alpert**, J. Hu*, S. Karandikar**, Z. Li*, W. Shi* and C. Sze** *Dept of ECE, Texas.
Outline Introduction: BTI Aging and AVS Signoff Problem
Inverter Chapter 5 The Inverter April 10, Inverter Objective of This Chapter  Use Inverter to know basic CMOS Circuits Operations  Watch for performance.
Optimal Placement of Energy Storage in Power Networks Christos Thrampoulidis Subhonmesh Bose and Babak Hassibi Joint work with 52 nd IEEE CDC December.
Logic synthesis flow Technology independent mapping –Two level or multilevel optimization to optimize a coarse metric related to area/delay Technology.
-1- UC San Diego / VLSI CAD Laboratory Optimization of Overdrive Signoff Tuck-Boon Chan, Andrew B. Kahng, Jiajia Li and Siddhartha Nath Tuck-Boon Chan,
1 Timing Closure and the constant delay paradigm Problem: (timing closure problem) It has been difficult to get a circuit that meets delay requirements.
-1- UC San Diego / VLSI CAD Laboratory Optimal Reliability-Constrained Overdrive Frequency Selection in Multicore Systems Andrew B. Kahng and Siddhartha.
1 Hardware Reliability Margining for the Dark Silicon Era Liangzhen Lai and Puneet Gupta Department of Electrical Engineering University of California,
PROCEED: Pareto Optimization-based Circuit-level Evaluation Methodology for Emerging Devices Shaodi Wang, Andrew Pan, Chi-On Chui and Puneet Gupta Department.
Unified Adaptivity Optimization of Clock and Logic Signals Shiyan Hu and Jiang Hu Dept of Electrical and Computer Engineering Texas A&M University.
THE CMOS INVERTER.
Standard-Cell Mapping Revisited
Circuit Design Techniques for Low Power DSPs
Sungho Kang Yonsei University
Reinventing The Wheel: Developing a New Standard-Cell Synthesis Flow
Fast Min-Register Retiming Through Binary Max-Flow
Presentation transcript:

Eyecharts: Constructive Benchmarking of Gate Sizing Heuristics Puneet Gupta, University of California, Los Angeles Andrew B. Kahng, University of California, San Diego Amarnath Kasibhatla, University of California, Los Angeles Puneet Sharma, Freescale Semiconductor, TX Research funded in part by NSF

Outline Motivation Solving Eyechart Topologies Experiments for Suboptimality Study Results 2

Why is Sizing Important? Sizing  Effective way of optimizing for power, speed and area  Tunable parameters  Gate-width  Threshold voltage  Gate-length  Supply voltage etc. Sizing problem seen at all stages of RTL to GDS flow Power recovery crucial during post-layout phase 3

Why Study Suboptimality? Literally hundreds of gate sizing methods exist Common heuristics/algorithms:  Linear Programming, Lagrangian Relaxation, Convex Optimization, Dynamic Programming, Geometric Programming, Sensitivity based gradient-descent, Simulated Annealing etc. Which heuristic is better? No systematic way to compare, so far How suboptimal are these heuristics? Does a heuristic’s performance depend on  Circuit topology?  Characteristics of the cell library? No prior work focuses on these aspects 4

Sizing Problem Formulation Sizing problem could be discrete or continuous Discrete sizing problem is NP-hard Common designs are standard cell-based and discrete We focus only on discrete sizing problem Problem: Leakage power minimization under timing constraints Circuits are purely combinational Gate sizing alone is tested and not logic optimization capability 5

Our Contributions We generate artificial combinational circuits called eyecharts Gate’s delay depends only on  Gate size  Total load capacitance Eyecharts can be solved optimally using Dynamic Programming (DP) A variety of eyecharts are generated by varying  Circuit topology  Power-Size, delay-size characteristics of library Suboptimalities of existing heuristics studied under these variations Leakage optimization details are presented Extensions to dynamic power optimization are easy 6

Outline Motivation Solving Eyechart Topologies Experiments for Suboptimality Study Results 7

What are Eyecharts? Chain MESH STAR Chain, Mesh and Star proposed Each can be solved optimally for the assumed delay model Stage of a gate is its logic level from the PI Levelized nature of eyecharts enables optimal sizing by DP 8

What are Eyecharts? Chain MESH STAR Stage1 Stage2 Stage3 Stage4 Stage5 PI PO Chain, Mesh and Star proposed Each can be solved optimally for the assumed delay model Stage of a gate is its logic level from the PI Levelized nature of eyecharts enables optimal sizing by DP 9

Solving a Chain Optimally InputLeakage Delay cappower Load Load = 3 = 6 Size Size

Solving a Chain Optimally Stage 3 Stage 1 Stage 2 6 INV1 INV2 INV3 D max = 8 InputLeakage Delay cappower Load Load 3 6 Size Size

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 Stage 3 Stage 1 Stage 2 ? InputLeakage Delay cappower Load Load 3 6 Size Size

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 Stage 1 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? Stage 3 Stage 1 Stage 2 Budget Power Size ? InputLeakage Delay cappower Load Load 3 6 Size Size

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 Stage 1 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? Stage 3 Stage 1 Stage 2 Budget Power Size ? Load = 3 InputLeakage Delay cappower Load Load 3 6 Size Size

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 Stage 1 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? Stage 3 Stage 1 Stage 2 Budget Power Size ? InputLeakage Delay cappower Load Load 3 6 Size Size Load = 6 15

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 Stage 1 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? Stage 3 Stage 1 Stage 2 Budget Power Size ? InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 16

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 Stage 1 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? Stage 3 Stage 1 Stage 2 Budget Power Size ? InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 17

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 Stage 1 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? Stage 3 Stage 1 Stage 2 Budget Power Size ? InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 18

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 Stage 1 1 ? ? 2 ? ? ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? Stage 3 Stage 1 Stage 2 Budget Power Size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 19

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 Stage Stage 3 Stage 1 Stage 2 Budget Power Size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 20

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 Stage Stage 1 Stage 3 Stage 1 Stage 2 Budget Power Size Budget Power Size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 Load = 3 Load = 6 21

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = Stage 1 Stage 2 Stage 3 Stage 1 Stage 2 Budget Power Size Budget Power Size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 Load = 3 Load = 6 22

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = Stage 1 Stage 2 Stage 3 Stage 1 Stage 2 Budget Power Size Budget Power Size INV2 Excess Total delay budget power size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 Load = 3 Load = 6 23

Solving a Chain Optimally 6 INV1 INV2 INV3 D max = 8 INV2 Excess Total delay budget power size size Stage 1 Stage 2 Stage 3 Stage 1 Stage 2 Budget Power Size Budget Power Size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 Load = 3 Load = 6 24

Solving a Chain Optimally INV1 INV2 INV3 D max = Stage 1 Stage 2 Stage 3 Stage 1 Stage 2 Budget Power Size Budget Power Size Budget Power Size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 Load = 3 Load = 6 Load = 6 25

Solving a Chain Optimally INV1 INV2 INV3 D max = Stage 1 Stage 2 Stage 3 Stage 1 Stage 2 Budget Power Size Budget Power Size Budget Power Size INV3 Excess Total delay budgetpower size size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 Load = 3 Load = 6 Load = 6 26

Solving a Chain Optimally INV1 INV2 INV3 D max = Stage 1 Stage 2 Stage 3 Stage 1 Stage 2 Budget Power Size Budget Power Size Budget Power Size INV3 Excess Total delay budgetpower size size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 Load = 3 Load = 6 Load = 6 27

Solving a Chain Optimally INV1 INV2 INV3 D max = INV3 Excess Total delay budgetpower size size Stage 1 Stage 2 Stage 3 Stage 1 Stage 2 Budget Power Size Budget Power Size Budget Power Size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 Load = 3 Load = 6 Load = 6 28

Solving a Chain Optimally INV1 INV2 INV3 D max = Stage 1 Stage 2 Stage 3 Stage 1 Stage 2 Budget Power Size Budget Power Size Budget Power Size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 Load = 3 Load = 6 Load = 6 29

Solving a Chain Optimally INV1 INV2 INV3 Load = 6 Budget Power Size 6fF OPTIMIZED CHAIN D max = 8 Stage 1 Stage 2 Stage 3 Stage 1 Stage 2 Budget Power Size Budget Power Size InputLeakage Delay cappower Load Load 3 6 Size Size Load = 3 Load = 6 Load = 3 Load = 6 30

Solving Mesh Optimally Stage with multiple gates represented with composite cell Mesh to chain conversion C2 C3 C4 A1 A2 B1 B2 Stage1 Stage2 Stage3 Stage4 Stage5 A1B2 A1 A2 A1 31

Solving Mesh Optimally Stage with multiple gates represented with composite cell Delay, power numbers for all size combinations for all output load combinations Delay, power table of B1 Mesh to chain conversion C2 C3 C4 A1 A2 B1 B2 Stage1 Stage2 Stage3 Stage4 Stage5 A1B2 InputPower Delay cap Load Load =12 = 24 Size Size Size Power Delay (B1,B2) (1,1)30 12 (1,2)506 (2,1)40 12 (2,2)604 Size Power Delay (B1,B2) (1,1)30 16 (1,2)50 8 (2,1)40 16 (2,2)60 8 LOAD = 12 LOAD = 24 STAGE 4 A1 A2 A1 32

Solving Star Optimally Star solved by converting it to chain Composite cells formed for stages with multiple gates A1 A2 B C1 C3 B Stage 1 Stage2Stage 3 C1 & C3, Composite cells for Stages 1 & 3 33

Hybrid Eyecharts Chain 1 and Chain 2 Chain 3 and Chain 4 Sample hybrid eyechart A A A A A B B B BA AA A C A A A A A B B B B A A A A A B B B B A A A A A B B B BA C Chain, mesh, star daisy-chained for arbitrarily large hybrid eyecharts Mesh/chain arbitrarily inserted along each PI/PO chain Hybrid eyechart solved optimally by ultimately reducing it to a chain A A A Chain 1 Chain 2 Chain 3 Chain 4 PO1 PO2 PI2 PI1 34

Arbitrary Extensions to Eyecharts Arbitrary extensions potentially add more realism to eyecharts Such topologies solved using partial enumeration No levelization restriction One example is multi-output mesh PO PI MESH 35

Arbitrary Extensions to Eyecharts Arbitrary extensions potentially add more realism to eyecharts Such topologies solved using partial enumeration No levelization restriction One example is multi-output mesh PO 1 PO 2 PO 3 PI MULTI-OUTPUT MESH PO PI MESH 36

Arbitrary Extensions to Eyecharts Arbitrary extensions potentially add more realism to eyecharts Such topologies solved using partial enumeration No levelization restriction One example is multi-output mesh PO 1 PO 2 PO 3 PI MULTI-OUTPUT MESH PO PI MESH 37

Outline Motivation Solving Eyechart Topologies Experiments for Suboptimality Study Results 38

Experimental Setup Heuristics compared  Comm1, Comm2: Two different commercial gate-sizing/leakage- optimization tools  GS: Sensitivity-based sizing tool with sensitivity metric =  LP: Linear programming tool [Nguyen et.al, ISLPED ’03]  SBS: Sensitivity-based sizing tool with sensitivity metric = [Gupta et.al, IEEE Tran. on CAD ’06] Explored power, delay tradeoffs with size  LP-LD: Linear increase in power, linear increase in delay  LP-NLD: Linear increase in power, nonlinear increase in delay  EP-LD: Exponential increase in power, linear increase in delay  EP-NLD: Exponential increase in power, nonlinear increase in delay Experiments to explore dependence of suboptimality on  Circuit size, circuit topology  Delay-Size, power-size tradeoff  Granularity of the cell library 39

Library Characteristics LP-LD 8.43% Gate Sizing 8 LP-NLD 0.3% Gate Sizing 8 EP-LD 8.43% V t, gate-length 3,3 EP-NLD 0.3% V t, gate-length 3,3 Library Model RMS fittingOptimizationDefault # error (delay) context sizes/variants The sizing choices for EP-LD and EP-NLD models are V t variants and gate-length variants Capacitance does not vary across V t variants Capacitance increase with gate-length for gate-length variants Delay values are fitted to a 65 nm industrial library Suboptimalities are calculated as 40

Outline Motivation Solving Eyechart Topologies Experiments for Suboptimality Study Results 41

Circuit Topology Impact Mesh-only Chain-only Suboptimality % Star-only Mesh is the toughest topology 42

Circuit Size Impact #Gates Comm1 RT Comm2 RT LP RT GS RT % % % % Suboptimality relatively constant with circuit size 10K-gate benchmarks for the rest of the experiments LP runtime does not scale well 43

Circuit Size Impact #Gates Comm1 RT Comm2 RT LP RT GS RT % % % % Suboptimality relatively constant with circuit size 10K-gate benchmarks for the rest of the experiments LP runtime does not scale well 44

Impact of Timing Constraints LP-LD delay model: Linear increase in delay with size Suboptmalities are close to zero for very tight or very relaxed constraints 45

Impact of Nonlinearity in Delay LP-NLD delay model: Nonlinear increase in delay with size Gap between sensitivity-based methods and LP tool becomes narrow 46

Impact of Power Tradeoff EP-NLD delay model: Exponential increase in power LP suffers significantly due to snapping error 47

Gate-length Biasing Scenario Tools in general slightly worse compared to V t assignment Capacitance varies across gate-length variants 48

Effect of Granularity Higher granularity has much larger benefits for exponential compared to linear power tradeoff LP-NLD EP-NLD 49

Extensions to Slew Dependent Delay Delay of a gate depends on input slew Output slew of a gate depends on  Gate size  Output load capacitance Slew propagation is not considered Optimal solution not guaranteed due to the need to maintain slew consistency Experiments show suboptimality is still significant (5% to 35%)

Conclusion Eyecharts can help diagnose and benchmark gate sizing heuristics  Suboptimality depends on circuit topology and power-/delay-size tradeoffs Existing heuristics be highly suboptimal  2% to 46% (gate sizing)  6% to 54% (V t -assignment)  14% to 49% (gate-length biasing) Ongoing work includes benchmarks  Based on fanout distribution  For joint sizing, multi-Vt, gate-length variant optimization Benchmarks and code can be downloaded at 50