Synopsys, Inc.

700 East Middlefield Road

Mountain View, CA 94043 USA

CCS Noise

Technical White Paper

Version 1.1

Abstract
This document describes the Synopsys CCS Noise model for cell-level noise
analysis.

Version 1.1

11/21/05
1
 Synopsys, Inc.
700 East Middlefield Road
Mountain View, CA 94043 USA

Copyright Notice and Proprietary Information

Copyright © 2005 Synopsys, Inc. All rights reserved. This software and documentation contain confidential and
proprietary information that is the property of Synopsys, Inc. The software and documentation are furnished under a
license agreement and may be used or copied only in accordance with the terms of the license agreement. No part of
the software and documentation may be reproduced, transmitted, or translated, in any form or by any means, electronic,
mechanical, manual, optical, or otherwise, without prior written permission of Synopsys, Inc., or as expressly provided by
the license agreement.
Right to Copy Documentation
The license agreement with Synopsys permits licensee to make copies of the documentation for its internal use only.
Each copy shall include all copyrights, trademarks, service marks, and proprietary rights notices, if any. Licensee must
assign sequential numbers to all copies. These copies shall contain the following legend on the cover page:“This
document is duplicated with the permission of Synopsys, Inc., for the exclusive use
of__________________________________________ and its employees. This is copy number __________.”
Destination Control Statement
All technical data contained in this publication is subject to the export control laws of the United States of America.
Disclosure to nationals of other countries contrary to United States law is prohibited. It is the reader’s responsibility to
determine the applicable regulations and to comply with them.
Disclaimer
SYNOPSYS, INC., AND ITS LICENSORS MAKE NO WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, WITH
REGARD TO THIS MATERIAL, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
Registered Trademarks (®)
Synopsys, AMPS, Arcadia, C Level Design, C2HDL, C2V, C2VHDL, Cadabra, Calaveras Algorithm, CATS, CSim,
Design Compiler, DesignPower, DesignWare, EPIC, Formality, HSPICE, Hypermodel, I, iN-Phase, in-Sync, Leda,
MAST, Meta, Meta-Software, ModelAccess, ModelTools, NanoSim, OpenVera, PathMill, Photolynx, Physical Compiler,
PowerMill, PrimeTime, RailMill, Raphael, RapidScript, Saber, SiVL, SNUG, SolvNet, Stream Driven Simulator, Superlog,
System Compiler, Testify, TetraMAX, TimeMill, TMA, VCS, Vera, and Virtual Stepper are registered trademarks of
Synopsys, Inc.
Trademarks (™)
abraCAD, abraMAP, Active Parasitics, AFGen, Apollo, Apollo II, Apollo-DPII, Apollo-GA, ApolloGAII, Astro, Astro-Rail,
Astro-Xtalk, Aurora, AvanTestchip, AvanWaves, BCView, Behavioral Compiler, BOA, BRT, Cedar, ChipPlanner, Circuit
Analysis, Columbia, Columbia-CE, Comet 3D, Cosmos, CosmosEnterprise, CosmosLE, CosmosScope, CosmosSE,
Cyclelink, Davinci, DC Expert, DC Expert Plus, DC Professional, DC Ultra, DC Ultra Plus, Design Advisor, Design
Analyzer, Design Vision, DesignerHDL, DesignTime, DFM-Workbench, DFT Compiler, Direct RTL, Direct Silicon
Access, Discovery, DW8051, DWPCI, Dynamic-Macromodeling, Dynamic Model Switcher, ECL Compiler, ECO
Compiler, EDAnavigator, Encore, Encore PQ, Evaccess, ExpressModel, Floorplan Manager, Formal Model Checker,
FoundryModel, FPGA Compiler II, FPGA Express, Frame Compiler, Galaxy, Gatran, HDL Advisor, HDL Compiler,
Hercules, Hercules-Explorer, Hercules-II, Hierarchical Optimization Technology, High Performance Option, HotPlace,
HSPICE-Link, iN-Tandem, Integrator, Interactive Waveform Viewer, i-Virtual Stepper, Jupiter, Jupiter-DP, JupiterXT,
JupiterXT-ASIC, JVXtreme, Liberty, Libra-Passport, Library Compiler, Libra-Visa, Magellan, Mars, Mars-Rail, Mars-Xtalk,
Medici, Metacapture, Metacircuit, Metamanager, Metamixsim, Milkyway, ModelSource, Module Compiler, MS-3200, MS-
3400, Nova Product Family, Nova-ExploreRTL, Nova-Trans, Nova-VeriLint, Nova-VHDLlint, Optimum Silicon, Orion_ec,
Parasitic View, Passport, Planet, Planet-PL, Planet-RTL, Polaris, Polaris-CBS, Polaris-MT, Power Compiler,
PowerCODE, PowerGate, ProFPGA, ProGen, Prospector, Protocol Compiler, PSMGen, Raphael-NES, RoadRunner,
RTL Analyzer, Saturn, ScanBand, Schematic Compiler, Scirocco, Scirocco-i, Shadow Debugger, Silicon Blueprint,
Silicon Early Access, SinglePass-SoC, Smart Extraction, SmartLicense, SmartModel Library, Softwire, Source-Level
Design, Star, Star-DC, Star-MS, Star-MTB, Star-Power, Star-Rail, Star-RC, Star-RCXT, Star-Sim, Star-SimXT, Star-
Time, Star-XP, SWIFT, Taurus, Taurus-Device, Taurus-Layout, Taurus-Lithography, Taurus-Process, Taurus-
Topography, Taurus-Visual, Taurus-Workbench, TimeSlice, TimeTracker, Timing Annotator, TopoPlace, TopoRoute,
Trace-On-Demand, True-Hspice, TSUPREM-4, TymeWare, VCS Express, VCSi, Venus, Verification Portal, VFormal,
VHDL Compiler, VHDL System Simulator, VirSim, and VMC are trademarks of Synopsys, Inc.
Service Marks (SM)
MAP-in, SVP Café, and TAP-in are service marks of Synopsys, Inc.
SystemC is a trademark of the Open SystemC Initiative and is used under license.
ARM and AMBA are registered trademarks of ARM Limited.
All other product or company names may be trademarks of their respective owners.
Printed in the U.S.A.

Version 1.1

11/21/05
2
 Synopsys, Inc.
700 East Middlefield Road
Mountain View, CA 94043 USA

1 Introduction
CCS Noise is a new advanced current-based driver model that enables
accurate noise analysis with results very close to SPICE simulation. It
not only precisely models injected crosstalk noise bumps, but also allows
more advanced analysis, such as propagated noise bumps and the driver
weakening, without significant characterization effort.
With CCS Noise, the noise immunity of the cells can also be obtained
during the analysis by using the actual noise bump waveforms without
the additional need for separate characterization. This dynamic
computation of noise propagation and noise immunity enables noise
library characterization to be 100 times faster than NLDM noise
characterization.
There are three main components to noise calculation: driver modeling,
receiver modeling, and reduced order modeling of parasitics. The focus
of this paper is the CCS Noise driver model; information about the
advanced receiver modeling can be found in the CCS Timing Technical
White Paper [1].

2 CCS Noise Driver Model

Current-based driver models are necessary for very accurate crosstalk
noise analysis. The desired cell current model for crosstalk noise analysis
must be able to interface with arbitrary input noise waveforms and
arbitrary coupled load interconnect networks.
CCS Noise is an advanced current driver model that captures both static
and transient characteristics of the cell. The static component of the CCS
Noise model consist of a current table as a function of input and output
voltage levels which can be provided through an efficient DC analysis
known as a basic ViVo model (Vin/Vout).
The key advantage of CCS Noise is that it uses several dynamic
parameters to model the dynamic response of the cell that a static
current table cannot capture. The dynamic parameters are extracted from
transient analysis measurements that record the response of the cell to
certain input transitions and noise bumps.
CCS Noise can accurately model all crosstalk noise analysis effects
including noise calculation, noise propagation, driver weakening, and
combined noise propagation and noise injection. Here, noise calculation
Version 1.1

11/21/05
3
 Synopsys, Inc.
700 East Middlefield Road
Mountain View, CA 94043 USA

refers to the problem of calculating an aggressor net’s injected noise

bump, assuming the victim driver is quiet. Noise propagation refers to the
problem of propagating a noise bump through a circuit cell, assuming
there is no coupling in the cell output net. Combined noise propagation
and noise injection analysis refer to the general case where there is noise
propagation through the victim driver and noise injections by aggressors
of the victim net. Driver weakening can be considered as a special case
of noise combination, where the propagated noise is very small by itself,
but significantly reduces the effective driving strength of the victim driver
due to increased injected noise bump size. Extensive studies have shown
that CCS Noise is considerably more accurate than other models for
these crosstalk noise analysis tasks, especially because of the CCS
Noise dynamic parameters. For example, Fig. 3 shows that propagated
noise waveforms computed using CCS Noise match SPICE waveforms
much better than those computed using the basic ViVo driver model
without the dynamic parameters.

(a) Basic ViVo model (b) CCS

NoiseCCS
Fig. 3. Waveform propagation comparison (Green curves are input waveforms; blue curves are
SPICE output waveforms, and dashed red curves are waveforms computed using models).

The CCS Noise analysis flow can take advantage of the advanced CCS
timing receiver model when such data are available in the library. More
details about the CCS timing receiver model can be found in the CCS
Timing White Paper [1]. For the purpose of noise analysis, the CCS
timing receiver model is more accurate than single-valued min/max
rise/fall pin capacitance because it includes the dependency of effective
receiver pin capacitance on input transition time and receiver output load
capacitance. Furthermore, the CCS Noise analysis engine explicitly
includes the varying effective input capacitance effect of victim receivers;
no extra receiver characterization is required.

Version 1.1

11/21/05
4
 Synopsys, Inc.
700 East Middlefield Road
Mountain View, CA 94043 USA

As discussed earlier, the ViVo-based current driver models best capture

the behavior of a single channel-connected block (CCB). For complex
circuit cells having more than one CCB, the transistor-level netlist of the
circuit cell needs to be broken into multiple CCBs and a CCS Noise
model for each of those CCBs. The transistor-level netlist breaking and
CCS Noise model parameter extraction are performed during cell
characterization.

Once characterized, CCS Noise model data are stored either on a timing
arc or on a pin in the cell library, depending on the topology of the circuit
netlist. For circuit cells having a single CCB for an input-output pin pair,
one CCS Noise model is extracted and stored on a timing arc. Such
single-stage cells include most inverters, NAND gates, NOR gates, AOI
gates, OAI gates, etc. For circuit cells having two subsequent CCBs, two
CCS Noise models are stored on a timing arc. Such two-stage cells
include most of the buffers, AND gates, OR gates, AND-OR gates, and
OR-AND gates, etc. For circuit cells having three or more CCBs,
including most of the flip-flops, full adders, macro blocks, etc., CCS
Noise model data are stored on pins.

3 CCS Noise Analysis

Cross-coupling between a victim net and aggressor nets causes a noise
bump on the victim net when the aggressors switch. As shown in Fig. 5,
when the noise bump is large enough and the propagated noise bump is
latched by the sequential cell, it will cause a functional error by changing
the logic value of the sequential cell. It will also increase the dynamic
power consumption induced by noise bumps. This issue can be
addressed by reporting potential violation sources in report_at_source
mode. In the report_at_source mode, any noise bump that fails the noise
immunity criteria is reported. You should fix all the reported violations to
address all the potential violation sources.

Version 1.1

11/21/05
5
 Synopsys, Inc.
700 East Middlefield Road
Mountain View, CA 94043 USA

injected noise

U1 U2 U3 FF
N1

possible functional error at FF

increased dynamic power consumption

Fig. 5. A violation reported at the report_at_source mode.

Alternatively, you could fix just those noise bumps that can propagate all
the way to an endpoint such as the D pin of a flip-flop; this method of
analysis is referred to as report_at_endpoint mode.

injected noise
attenuated noise

U1 U2 U4 FF
N1
highly immune cell

Fig. 6. The injected noise attenuates when it propagates.

As an example, consider the scenario shown in Fig. 6, where U4 is a

highly noise immune cell; that is, it can tolerate a certain amount of noise
without causing a failure. Its propagated noise is also significantly
attenuated such that no functional error occurs in FF. In the
report_at_source mode, the input pin of U2 would be reported as a
violation because the noise bump exceeds the noise immunity limit.
However, in the report_at_endpoint mode, no violation will be reported
because the injected noise bump attenuates in a later stage and does not
cause any functional error at the endpoint. Therefore, when users care
only about the failures at the inputs of sequential cells, the
Version 1.1

11/21/05
6
 Synopsys, Inc.
700 East Middlefield Road
Mountain View, CA 94043 USA

report_at_endpoint mode analysis can potentially report fewer noise

violations .

Fig. 7 shows an example of the report_at_endpoint mode analysis. In this

figure, a source of violation is the victim net N1, where the noise bump is
created due to the cross coupling. A noise endpoint is any input pin of the
sequential cell where the noise propagation stops; in this case, the input
pin of FF

propagated noises
injected noise

U1 U2 U3 FF
N1

violation endpoint
violation source
Fig. 7. Source of violation vs. violation at an endpoint.

In the report_at_endpoint mode, it may be difficult to determine the

violation sources just by viewing the reported violations at noise
endpoints, such as the input pins of sequential cells. But the violation
sources can be found by searching backward through the fan-in logic
cones from the violation endpoints, as shown in Figure 8.

Version 1.1

11/21/05
7
 Synopsys, Inc.
700 East Middlefield Road
Mountain View, CA 94043 USA

U1

U2 U4 FF
N1

U3
N2
backward searching for sources

Fig. 8. Reporting the sources of violations in the report_at_endpoint mode analysis.

Fig. 8 shows how the violation sources can be found given the violation
reported at the endpoint. The backward search starts from the endpoint,
and traces the fan-in logic cone until it hits the first net where the injected
noise exceeds the noise immunity. In this way, all possible sources of
violations can be found. In Figure 8, the violation has occurred at the D
pin of FF and report_at_endpoint mode reports nets N1 and N2 as the
violation sources. So when you fix the violations on N1 and N2, the
violation at the endpoint will be fixed as well.

4 Noise Slack Calculation

A CMOS circuit cell can tolerate a certain amount of noise without
causing a failure at the cell output; this characteristic is called noise
immunity. CCS Noise information can be used to calculate the noise
immunity of a cell on the fly.

Using CCS Noise and the actual shape of the input noise bump, the
noise immunity of the cell can be computed as well as the noise slack;
Version 1.1

11/21/05
8
 Synopsys, Inc.
700 East Middlefield Road
Mountain View, CA 94043 USA

that is, the amount of noise height that needs to be added to the noise
bump to cause a failure. When the noise slack is negative, it means that
the noise bump has exceeded the noise immunity parameters for the cell.
When the noise slack is positive, the noise bump is within the noise
immunity parameters.

5 Results
Fig. 9 shows the results of a crosstalk noise correlation with SPICE. The
test circuit has two aggressor nets coupled to a victim net and the victim
net has a fan-out net for noise propagation analysis.
Fig. 9 (b) compares CCS Noise and SPICE waveforms for a typical case.
The difference in waveforms is barely discernable at the resolution of the
plot. The cell driving strengths: that is, the input transition times of the
aggressor drivers and the lengths of interconnects, have also been varied
to cover all reasonable scenarios for accuracy correlation.
The noise bump height correlation at point A (for noise calculation) is
shown in Fig. 9 (c). The correlation at point B (for both noise calculation
and propagation) is shown in Fig. 9 (d).
aggressor

victim A B

aggressor (a) test circuit

A B

(b) comparison of waveforms (c) correlation at A (d) correlation at B

Fig. 9. Comparison of CCS NoiseCCS NoiseCCS NoiseCCS Noise analysis results with SPICE.

Version 1.1

11/21/05
9
 Synopsys, Inc.
700 East Middlefield Road
Mountain View, CA 94043 USA

Note that a coupled interconnect network for crosstalk noise analysis is a

nonlinear system, therefore the law of linear superposition does not
apply. CCS Noise enables the combined noise bump by switching all
aggressors. The sum of individual noise bumps may not match the total
noise bump.

Table 1 and Fig. 10 show details of static noise analysis results with
customer designs.

Table 1 shows how the pessimism is reduced in different noise analysis

modes by counting the number of violations reported. The fourth column
shows the number of noise bumps whose heights are higher than 40% of
VDD. However, after the noise immunity and noise slack for those noise
bumps are calculated, no violations are reported because the noise
bumps do not exceed their noise immunity. Only the noise bumps whose
noise slacks are negative are reported in the report_at_source column.
Moreover, if you exclude any noise bumps that attenuate during their
propagation and instead report only the noise bumps that reach the
endpoints and cause functional errors, pessimism can be further reduced.
These numbers are reported in the report_at_endpoint column. As
shown in the table, report_at_endpoint mode analysis reports fewer
violations compared with the number of noise bumps with their heights
greater than 40% of VDD. When the report_at_endpoint mode analysis is
used, this pessimism reduction can save lots of time during ECO fixing
stage, and significantly reduces the number of ECO fixing iterations.

Table 1. Number of static noise violations in customer designs.

Number of violations
Number
Design Library height > 40%
of gates report_at_source report_at_endpoint
VDD
Design A 325K 90 nm 288 31 4
Design B 71K 90 nm 353 51 12
Design C 195K 90 nm 161 45 1

Version 1.1

11/21/05
10
 Synopsys, Inc.
700 East Middlefield Road
Mountain View, CA 94043 USA

400

300

violations
200

100

0
Design A Design B Design C

height > 40% VDD report_at_source report_at_endpoint

Fig. 10. Comparison of noise violation counts.

6 Summary
In this paper we have introduced the new advanced CCS Noise model
and shown how it can be used to improve static noise analysis. CCS
Noise provides for accurate noise calculation including the effects of
noise propagation, so that the gate model can be characterized
significantly faster than before. Results with various designs show that
CCS Noise matches SPICE simulation results with excellent correlation.

7 References
[1] CCS Timing Technical White Paper.

Version 1.1

11/21/05
11