Title: On Design-Manufacturing Integration Talk at Intel June 25, 2003 Andrew B. Kahng, UCSD CSE
1On Design-Manufacturing IntegrationTalk at
IntelJune 25, 2003Andrew B. Kahng, UCSD CSE
ECE Departmentsemail abk_at_ucsd.eduURL
http//vlsicad.ucsd.edu
2Outline
- The Problem, Scope and Goals
- Example Cost-Driven RET
- Example Intelligent MDP
- Example Analog Rules, Restricted Layout,
- Conclusions
3The Problem
- Steadily increasing design effort, turnaround
time, and project risk - Example symptoms (next slides) detailed routing
- Dark Future (12th Japan DA Show talk, 2000)
- Cost and predictability failures ?
- Electronics industry makes workarounds
- platforms ? programmability ? software ?
- Semiconductor industry stalls ?
- No retooling cycle for supplier industries (e.g.,
EDA) - Cf. Intel announcement re 157nm lithography
4Reticle Enhancement (vs. Routing)
- 193nm, 157nm are late ? ?, radius of influence gtgt
F - Pattern context increasingly dominant
- Iso vs. dense, microloading/halation,
- RETs knobs aperture, phase, light, context
- Affect feasibility of mask write, verification
- ? Complexity trend model-based RETs, AIM tools
in the loop, test structure-based RLCX
calibration, etc.
5Routing Rules (1)
- Minimum area rules and via stacking
- Stacking vias through multiple layers can cause
minimum area violations (alignment tolerances,
etc.) - Via cells can be created that have more metal
than minimum via overlap (used for intermediate
layers in stacked vias) - Multiple-cut vias
- Use multiple-cut vias cells to increase yield and
reliability - Can be required for wires of certain widths
- Multiple via cut patterns have different spacing
rules - Four cuts in quadrilateral five cuts in cross
six cuts in 2x3 array - With wide-wire spacing rules, complicates pin
access - Cut-to-cut spacing rules ? check both cut-to-cut
and metal-to-metal when considering via-to-via
spacing - Line-end extensions
- Vias or line ends need additional metal overlap
(0th-order OPC)
6Routing Rules (2)
- Width- and Length-dependent spacing rules
- Width-dependent rules domino effects
- Variant parallel-run rule (longer parallel
runs ? more spacing) - Measuring length and width halo rules affect
computation - Influence rules or stub rules
- A fat wire, e.g., power/ground net, will
influence the spacing rule within its
surroundings ? any wire that is X um away from
the fat wire needs to be at least Y um away from
any other geometry. - Example fat wire with thin tributaries
- bigger spacing around every wire within certain
distance of the thin tributaries - ECO insertion of a tributary causes complications
- Strange jogs and spreading when wires enter an
influenced area
7Example LEF/DEF 5.5, April 2003
8Example LEF/DEF 5.5, April 2003
9Routing Rules (3)
- Density
- Grounded metal fills (dummy fill)
- Via isodensity rules and via farm rules (via
layers must be filled and slotted, have
width-dependent spacing rule analogs, etc.) - Non-rectilinear (?-geometry) routing
- X-Architecture http//www.xinitiative.org/
- Y-Architecture http//vlsicad.ucsd.edu/Yarchitec
ture/ , LSI Logic patents - Landing pad shapes (isothetic rectangle vs.
octagon vs. circle), different spacings (1.1x)
between diagonal and Manhattan wires, etc. - More exceptions
- More non-default classes (timing, EM reliability,
) - Not just power and clock
- gt0.25um width may be wide ? many exceptions
10Routing Rules
- Degrade completion rates, runtime efficiency
- Postprocessing likely no longer suffices
- E.g., antennas
- Can (should) (must) routers natively address
these issues?
11Other Symptoms NRE Cost
- Runtimes for mask data prep (MDP), mask write,
mask inspection - NRE cost ? runtimes ? shapes complexity
Context-dependent fracturing
P. Buck, Dupont Photomasks, July 2001
12Other Symptoms BEOL Yield
- More than half of catastrophic yield loss is in
BEOL - Copper is deposited ? can infer yield loss
mechanisms - Open faults (3x more prevalent than short or
bridging faults) - High-resistance via faults
- Cf. non-tree routing for reliability and yield?
- Huge amount of variability budget is in
planarization - Copper is soft ? dual-material polish mechanisms
- Oxide erosion and copper dishing ?
cross-sectional variability, inter-layer bridging
faults, - Much, much more
- Low-k dielectrics thermal properties,
anisotropy, nonuniformity - Resistivity at small conductor dimensions (grain
and barrier layer effects)
13Problem Charging and Antennas
- Process steps use plasmas, charged particles
- Charge collects on conducting poly, metal
surfaces - Capacitive coupling large electrical fields
over gate oxides cause damage or complete
breakdown - Induced Vt shifts affect device matching (e.g.,
for analog) and timing predictability - Solution limit antenna ratio (Apoly AM1 )
/ Agate-ox - AMx metal(x) area that is electrically
connected to node without using metal (x1), and
not connected to an active area - E.g., antenna ratio 400 for thicker oxide, 2000
for thin oxide - Two antenna ratios thin-ox (core) and thick-ox
(IO) - Several antenna rules per-layer, cumulative,
cap-based
14Antenna Management and Futures
- Bridging (break antenna by hopping to higher
layer) - Extra wiring, vias, congestion
- Antenna ratios (and gate areas) decrease ? more
bridges - Reverse-biased diode or source-drain contact near
gate - Leakage, area, timing penalties
- Futures
- Will antenna ratios decrease (high-k gate
dielectrics ? increased physical Tox ? less
leaky, hard failure modes)? - More preemption (no more post-processing, or
dioded cells)? - Explicit tradeoff of unfixed antenna yield
penalty for fixed antenna yield loss?
15Problem Layout Density Control
- Area fill electrically inactive floating or
grounded - Area fill insertion (and, slotting)
- Decreases local density variation
- ? Decreases post-CMP ILD erosion, conductor
dishing - Cf. Filling and Slotting Analysis and
Algorithms, ISPD-98
Post-CMP ILD thickness
Features
Area fill features
Cao et al. U. Wisconson
16Density Management and Futures
- Physical model based density rules and fill
synthesis - Current coevolved state Wrong rules, weak
tools
W. Grobman, Motorola, DAC-2001
17Evolutionary Paths
- Conflicting goals
- Designer freedom, reuse, migration
- EDA maintenance mode
- Process/foundry enhance perceived value
( add rules) - ? Prisoners Dilemma
- Fiddling Incremental, linear extrapolation of
current trajectory - GDS-3
- Thin post-processing layers (decompaction, RET
insertion, )
18DAC-2003 Nanometer Futures PanelWhere should
extra RD be spent?
19Co-Evolutionary Paths
- Designer, EDA, and process communities cooperate
and co-evolve to maintain the cost (value)
trajectory of Moores Law - Must escape Prisoners Dilemma
- Must be financially viable
- At 90nm to 65nm transition, this is a matter of
survival for the worldwide semiconductor industry - Example Focus Areas
- Manufacturability and cost/value optimization
- Restricted layout
- Intelligent mask data prep
- Analog rules
- (Layout and design optimizations)
- Disclaimer Not a complete listing
20Basic Goals
- Bidirectional design-manufacturing data pipe
- Fundamental drivers cost, value
- Pass functional intent to mask flow
- Example RET for predictable circuit
performance, function - RETs should win , reduce performance variation
- ? cost-driven, parametric yield constrained RET
- Pass limits of mask flow up to design
- Example avoid corrections that cannot be
manufactured or verified - N.B. 1998-2003 papers/tutorials
http//vlsicad.ucsd.edu/abk/TALKS/
21Outline
- The Problem, Scope and Goals
- Example 1 Performance-Driven Fill
- Example 2 Cost-Driven RET
- Example 3 Intelligent MDP
- Example 4 Analog Rules, Restricted Layout,
- Conclusions
22Layout Density Control
- Area fill electrically inactive, floating or
grounded - Area fill insertion (and slotting)
- Decreases local density variation
- ? Decreases post-CMP ILD erosion, conductor
dishing - Cf. Filling and Slotting Analysis and
Algorithms, ISPD-98
Post-CMP ILD thickness
Features
Area fill features
23Timing, Parametric Yield Impact
- Performance Impact Limited Fill (PIL-Fill),
DAC-2003 - Fill adds capacitance, hurts timing and SI
closure - Plain capacitance minimization objective is not
sufficient - CMP modeling ? layout density vs. dimensions
built into RLCX
1
top view
Active lines
2
B
A
C
Active lines
3
w
fill grid pitch
D
E
4
5
buffer distance
F
G
6
24Min-Slack, Fill-Constrained PIL-Fill
- Inputs LEF/DEF, extracted RSPF, STA (slack)
report - Drive ILP and greedy PIL-Fill methods by
estimated lateral coupling and Elmore delay
impact - Baseline comparison LP/Monte-Carlo methods
- Iterated greedy method for MSFC PIL-Fill reduces
timing slack impact of fill by 80 (average over
all nets), 63 (worst net)
25Other Density Management DOFs
- Splitting for uniformity
- Slotting for uniform CMP replaced by splitting
- Less data than traditional slotting
- Power mesh more accurate R/C analysis
- Combined density control and local pattern
control (e.g., fill helps iso-dense, PSM
phase-assignability) - Details hierarchy, reuse, multiple length
scales,
Easy connections through standard via arrays
GND
GND
GND
GND
VS.
M1
M1
Difficult to connect - where should vias go?
Illustration courtesy Cadence Design Systems, Inc.
26Outline
- The Problem, Scope and Goals
- Example Cost-Driven RET
- Example Intelligent MDP
- Example Analog Rules, Restricted Layout,
- Conclusions
27Design for Value
- Mask cost trend ? Design for Value (DFV)
- Design for Value Problem
- Given
- Performance measure f
- Value function v(f)
- Selling points fi corresponding to various values
of f - Yield function y(f)
- Maximize Total Design Value ?i y(fi)v(fi)
- or, Minimize Total Cost
- Probabilistic optimization regime
- See "Design Sensitivities to Variability
Extrapolation and Assessments in Nanometer VLSI",
IEEE ASIC/SoC Conference, September 2002, pp.
411-415.
28Obvious Step Function-Aware OPC
- Annotate features with required amount of OPC
- E.g., why correct dummy fill?
- Determined by design properties such as setup and
hold timing slacks, parametric yield criticality
of devices and features - Reduce total OPC inserted (e.g., SRAF usage)
- Decreased physical verification runtime, data
volume - Decreased mask cost resulting from fewer features
- Supported in data formats (OASIS, IBM GL-I)
- Design through mask tools need to make, use
annotations
29Cost-Driven RET
- MinCorr (DAC-2003) Different levels of RET
different levels of CD control
Type of OPC Ldrawn (nm) 3? of Ldrawn Figure Count Delay (?, ?) for NAND2X1
Aggressive 130 5 5X (60.7, 7.03)
Medium 130 6.5 4X (60.7, 7.47)
No OPC 130 10 1X (60.7, 8.79)
OPC solutions due to K. Wampler, MaskTools,
March 2003
CD studies due to D. Pramanik, Numerical
Technologies, December 2002
30Performance Measure Delay
- Selling point delay circuit delay which
achieves desired level (say 99) of parametric
yield - Goal Achieve selling point delay with minimum
cost of RETs (OPC)
31MinCorr The Cost of Correction DFV Problem
- Given Admissible levels of correction for each
layout feature and the corresponding delay impact
(mean and variance) - Find Level of correction for each layout
feature such that a prescribed selling point
delay is attained - Objective Minimize total cost of corrections
32Statistical Timing Analysis
Statistical STA (SSTA) provides PDFs of arrival
times at all nodes
33Variation Aware Library Model
- Capacitance and delay values replaced by (?,?)
pair - Sample variation aware .lib
- pin(A)
- direction input
- capacitance (0.002361,0.0003)
-
-
- timing()
- related_pin "A"
- timing_sense positive_unate
- cell_rise(delay_template_7x7)
- index_1 ("0.028, 0.044, 0.076")
- index_2 ("0.00158, 0.004108, 0.00948")
- values ( \
- (0.04918,0.001), (0.05482,0.0015),
(0.06499,0.002)", - .
34Generic Cost of Correction Methodology
Nominally Correct SPR Netlist
Min. Corrected Library
- Statistical STA (SSTA) provides PDFs of arrival
times at all nodes - Assume variation aware library models (for delay)
are available
SSTA
Yield Target met ?
Y
EXIT
N
All Correction Libraries
Correction Algorithm
All Correction Libraries
SSTA
35Generic Cost of Correction Methodology
Nominally Correct SPR Netlist
Min. Corrected Library
- Statistical STA (SSTA) provides PDFs of arrival
times at all nodes - Assume variation aware library models (for delay)
are available - Statistical STA currently has runtime and
scalability issues
SSTA
Yield Target met ?
Y
EXIT
N
All Correction Libraries
Correction Algorithm
All Correction Libraries
SSTA
36MinCorr Parallels to Gate Sizing
- Assume
- Gaussian-ness of distributions prevails
- ? 3? corresponds to 99 yield
- Perfect correlation of variation along all paths
- Die-to-Die variation
- ?12 3?12 ?1 3?1 ?2 3?2
- Resulting linearity allows propagation of (?3?)
or 99 (selling point) delay to primary outputs
using standard Static Timing Analysis (STA) tools
37MinCorr Parallels to Gate Sizing
Gate Sizing Problem Given allowed areas and
corresponding delays of each cell, minimize total
die area subject to a cycle time constraint
Gate Sizing ? MinCorr
Cell Area ? Cost of correction
Nominal Delay ? Delay (?k?)
Cycle Time ? Selling point delay
Die Area ? Total cost of OPC
38Components of MinCorr Sizing
- A yield-aware library that captures
- Delay mean and variance for each library master
for each level of correction - Relative cost of OPC for each master
corresponding to each level of correction - Use standard off-the-shelf logic synthesis tool
to perform sizing - Can use well-tested sizing methods
- Practical runtimes
- Can handle interesting variants, e.g.,
cost-constrained selling point delay minimization
39MinCorr Yield Aware Library Characterization
- Mask cost is assumed proportional to number of
layout features - Monte-Carlo simulations, coupled with linear
interpolation, are used to estimate delay
variance given the CD variation - We generate a library similar to Synopsys .lib
with (?3?) delay values for various output loads - Cost modeled by relative figure count multiplied
by the number of transistors in the cell - Gate input capacitance variation with CD
considered
40Experiments and Results
- Synopsys Design Compiler used as the synthesis
tool to perform gate sizing - Figure counts, critical dimension (CD)
variations derived from Numerical Technologies
OPC tool - Use a restricted TSMC 0.13 ?m library
- 7 cell masters BUF, INV, NAND, NOR
- Approach tested on small combinational circuits
- alu128 8064 cells
- c7552 2081 cell ISCAS85 circuit
- c6288 2769 cell ISCAS85 circuit
- Courtesy Dipu Pramanik, NTI
41Yield Library Generation
Type of OPC Ldrawn (nm) 3? of Ldrawn Figure Count Delay (?, 3?) for NAND2X2
Aggressive 130 5 5X (64.82, 2.14)
Medium 130 6.5 4X (64.82, 2.80)
No OPC 130 10 1X (64.82, 4.33)
Sample Result of Library Generation
- Three levels of OPC considered
- Input slew dependence ignored
- Interconnect variation ignored
42Cost Savings with MinCorr Sizing
Design Normalized Cost Normalized Selling Point Delay
alu128 5.0 (Aggressive OPC) 0.9644
4.0 (Medium OPC) 0.9739
1.0 (No OPC) 1.0000
1.0657 0.9644
1.0119 0.9976
c7552 5.0 0.9432
4.0 0.9621
1.0 1.0000
1.4639 0.9432
1.2079 0.9848
c6288 5.0 0.9480
4.0 0.9642
1.0 1.0000
4.1530 0.9480
43Cost Savings with MinCorr Sizing
- Small (5) selling point delay variation between
max- and min-corrected versions of design (5X
difference in cost) - Sizing-based optimization achieves 17-79
reduction in OPC cost without sacrificing
parametric yield
44Outline
- The Problem, Scope and Goals
- Example Cost-Driven RET
- Example Intelligent MDP
- Example Analog Rules, Restricted Layout,
- Conclusions
45MDP for Minimum NRE
- Mask data prep (MDP)
- Partition layout shapes into multiple gray-scale
writing passes - Determine apertures, beam currents, dwell times,
shot ordering, - Write error X MEEF gives wafer CD error
- Examples
- Simplified write of diagonal lines when
triangular aperture is available - Raster vs. Vector write
- Major field layout, subfield scheduling,
- Driven by performance analyses, sensitivities,
costs - Many other manufacturing NRE optimizations
available
46Example Triangular Apertures
www.xinitiative.org
47Subfield Scheduling (SPIE Microlithography03)
- Use of high-energy e-beams in mask writing is
limited by resist heating effects (resist
distortion, irreversible chemical changes) - Corrective measures (lower beam current density,
delays between flashes, multi-pass writing,
etc.) reduce throughput
Mask Writing Schedule Problem Given Beam
voltage, resist parameters, mask write throughput
requirement Find Beam current density and
subfield writing schedule to minimize the maximum
resist temperature Tmax
48Throughput-Normalized Subfield Scheduling
49Outline
- The Problem, Scope and Goals
- Example Cost-Driven RET
- Example Intelligent MDP
- Example Analog Rules, Restricted Layout,
- Conclusions
50Analog Rules
- We dont need no ((! rules
- Rules just make lithographers feel better (?)
- Ultimately, bottom line is cost of ownership,
TCOG - Given adequate models of MDP, RET and Litho
flows, design tools can and should optimize
parametric yield, /wafer, profits - More examples critical-area reduction by
decompaction, introducing redundancy (vias,
wires), - Automated learning of models and implicit rules
- Current approach test wafers, test structures,
second-hand understanding - S. Teig, ISPD-2002 keynote machine learning
techniques
51Restricted Layout
- Soft reset 1-time hit on Moores Law density
scaling - Restricted Design Rules (RDR) can be
compensated many ways - embedded 1-T SRAM fabric, stacking, I/O circuit
design, - N.B. Moores Law is a meta Law!
Dual Exposure Result
Islands
Checkerboard
Example PhasePhirst! (Levenson et al.)
0?
180?
Transparent
Opaque
Trim Mask Exposure
First Exposure
Dark-Field PSMs
or
M. D. Levenson, 2003
52Pre-Made Mask Blanks
- PhasePhirst!
- Levenson/Petersen/Gerold/Mack, BACUS-2000
- Idea mask blanks can be mass-produced and stored
to reduce average cost (e.g., 10K), then
customized on demand
M. D. Levenson, 2003
53Notes on Regular Layout
- 65 nm has high likelihood for layouts to look
like regular gratings - Uniform pitch and width on metal as well as poly
layers - ? Predictable layouts even in presence of focus
and dose variations - More manufacturable cell libraries with regular
structures - New layout challenges (e.g., preserving
regularity in placement) - Caveats
- Non-minimum width poly is less susceptible to
variation ? larger devices (with same W/L) may
lead to more robust designs (so, selective
upsizing may be an alternative to regular
layouts) - 180nm node is a sweet spot for cost,
analog/mixed-signal integration, etc. ? more and
more technology nodes will tend to coexist
54Outline
- The Problem, Scope and Goals
- Example Cost-Driven RET
- Example Intelligent MDP
- Example Analog Rules, Restricted Layout,
- Conclusions
55Conclusions
- Designer, EDA, and mask communities must
cooperate and co-evolve to maintain the cost
(value) trajectory of Moores Law - Wakeup call for supplier industries Intel
skipping 157nm litho - Basic goal bidirectional design-mask data pipe
- Drivers cost, value
- Pass functional intent to mask flow
- Pass limits of mask flow up to design
- Example focus areas (not a complete listing!)
- Manufacturability and cost/value optimization
- Restricted layout
- Intelligent mask data prep
- Analog rules
- (Layout and design optimizations)