A BUS ENCODING TO REDUCE CROSSTALK NOISE EFFECT IN SYSTEM ON CHIP

International Journal of VLSI design & Communication Systems (VLSICS) Vol.2, No.2, June 2011
DOI : 10.5121/vlsic.2011.2209 110
A BUS ENCODING TO REDUCE CROSSTALK NOISE
EFFECT IN SYSTEM ON CHIP
J.Venkateswara Rao1
and A.V.N.Tilak2
1
Department of Electronics and Communication Engineering, Vignan Institute of
Technology & Science, Deshmukhi, Nalgonda Dist. A.P, India.
jvrao_07@yahoo.com
21
Department of Electronics and Communication Engineering, Gudlavalleru
Engineering College Gudlavalleru, Krishna Dist. A.P, India.
avntilak@yahoo.com
ABSTRACT
This paper proposes a new bus coding scheme for reducing the crosstalk in System on chip(soc). As
circuit geometries become smaller, wire interconnections become closer together and taller, thus
increasing the cross-coupling capacitance between nets. At the same time, parasitic capacitance to the
substrate becomes less as interconnections become narrower, and cell delays are reduced as transistors
become smaller. With circuit geometries at 0.25 micron and above, substrate capacitance is usually the
dominant effect. However, with geometries at 0.18 micron and below, the coupling capacitance between
nets becomes significant, making crosstalk analysis increasingly important for accurate timing analysis.
We show experimentally that the proposed codes allow reducing crosstalk delay by at least 14% based on
available data.
KEYWORDS
Crosstalk, Encoding, SOC, parasitic, coupling Capacitance, micron, Forbidden Pattern free
1. INTRODUCTION
As device geometries shrink, chip sizes increase, and clock speeds get faster, interconnect delay
is becoming increasingly significant. Signal integrity is the ability of an electrical signal to carry
information reliably and resist the effects of high-frequency electromagnetic interference from
nearby signals. Crosstalk is the undesirable electrical interaction between two or more
physically adjacent nets due to capacitive cross-coupling. As integrated circuit technologies
advance toward smaller geometries, crosstalk effects become increasingly important compared
to cell delays and net delays.
In particular, the propagation delay through long cross-chip buses is already proving to be a
limiting factor in the speed of some designs, and this trend will only get worse. It has been
shown that the delay through a long bus is strongly a function of the coupling capacitance
between the wires. Especially detrimental to the delay is the Miller-like effect when adjacent
wires simultaneously transition in opposite directions. When the cross-coupling capacitance is
comparable to or exceeds the loading capacitance on the wires, the delay of such a transition
may be twice or more that of a wire transitioning next to a steady signal. We call this delay
penalty the “crosstalk delay”. In some high-speed designs where crosstalk delay would have
limited the clock speed, the technique of shielding was used. This involves putting a grounded
wire between every signal wire on the bus. Although this certainly is effective in preventing
crosstalk within the bus, it has the effect of doubling the wiring area. Cross-chip buses often
must be routed in higher metal layers, which are scaled more slowly than the rest of the

111
geometry in order to prevent an unacceptable increase in resistance. Thus, routing resources are
scarce at these levels, and it can be difficult to justify doubling the bus width.
However, if we abstract the concept of shielding and just look at the signals on the wires of a
shielded bus, we think of it as a very simple bus encoding. Two wires are used for every data
bit. A data bit of “0” is encoded as a “00” signal on the wires, and a “1” is encoded as “l0”.
The purpose of this “encoding” is to prevent adjacent wires from transitioning in opposite
directions, and this particular encoding achieves that goal by forcing every other wire to a
steady value. But the question arises: Are there other possible encodings that can achieve the
same goal, but with fewer wires? Such encodings may require extra logic or memory elements,
but as the speed of logic goes up and the relative area consumed by logic goes down, such a
tradeoff seems increasingly valid. Indeed, such encodings exist. We will refer to them as “self-
shielding” or “crosstalk-immune” codes. In this paper, we designed the codes based on the
codes theoretical suggested in [7]. We will develop the theory behind crosstalk-immune coding,
describe the fundamental capabilities and limitations that the theory implies, and give methods
for generating optimal sets of code words.
The most aggressive of our encoding techniques actually speeds up a bus by exploiting
crosstalk. In this encoding approach, if a bus signal rises (falls), then our encoding forces one of
its neighbors to rise (fall) as well, while the other neighbor is static. As a result, we actually
improve the rise time of the wire by utilizing crosstalk to our best. This is not possible with
current approaches to alleviate the cross-talk problem in buses (which stagger bus signals in
space and/or time to mitigate the cross-talk problem among bus signals). In recent times, with
wiring delays increasing compared to gate delays [8], it is often the case that the critical delay in
a circuit is determined by long buses. In such a case, buses could be encoded with the
techniques described in this paper, allowing the design to be operated at a much greater
frequency. A designer would gladly tolerate the bus size overhead involved with the use of our
approach, in such a scenario.
Different approaches have been proposed for crosstalk reduction in the context of bus
interconnects. Some schemes focus on reducing the energy consumption, some focus on
minimizing the delay and other schemes address both. The simplest approach to address the
inter-wire crosstalk problem is to shield each signal using grounded conductors. Khatri et al. in
[12] proposed a layout fabric that alternatively inserts one ground wire and one power wire
between every signal wire, i.e., the wires are laid out as . . . VSGSVSGSVS . . . , where S denotes
a signal wire, G denotes a ground wire and V denotes a power wire. Compared to passive
shielding, active shielding is a more aggressive technique that reduces the bus delay by up to
75% [13]. Ghoneima and Ismail in [14] investigated the theoretical optimum position for
repeater insertions to achieve minimum delay. Again, this technique is a physical design
technique and requires careful trace layout and repeater placement. The average delay can also
be reduced using statistical approaches such as data packing and data permutation proposed by
Satyanarayana et al. in [15]. All these techniques exploit the temporal coherency of bus data to
minimize the average delay for data transmission.
This paper is organized as follows. Section 2 provides crosstalk delay effects. In Section 3, we
discuss previously published approaches for solving this problem. In Section 4, we describe our
approach of creating memoryless-based crosstalk canceling CODECs. In Section 5 we report
the results of experiments that we have performed to quantify the tradeoff between the degree of
crosstalk immunity achieved by the above techniques, and the bus size overhead incurred. In

112
section 6, we compare our bus size overheads with those reported in [7], in which memoryless
CODECs to eliminate crosstalk patterns were described. We conclude the paper in Section 7.
2. PRELIMINARIES OF CROSSTALK EFFECTS
Crosstalk can affect signal delays by changing the times at which signal transitions occur. For
example, consider the signal waveforms on the cross-coupled nets A, B, and C in fig. 1.
Because of capacitive cross-coupling, the transitions on net A and net C can affect the time at
which the transition occurs on net B. A rising-edge transition on net A at the time shown in Fig.
1 can cause the transition to occur later on net B, possibly contributing to a setup violation for a
path containing B. Similarly, a falling-edge transition on net C can cause the transition to occur
earlier on net B, possibly contributing to a hold violation for a path containing B.
Figure 1. Transition Slowdown or Speedup Caused by Crosstalk.
Figure 2. Glitch Due to Crosstalk.
The logic effects of crosstalk on steady-state nets are due to the cross-coupled nets as shown in
fig. 2. Net B should be constant at logic zero, but the rising edge on net A causes a noise bump
or glitch on net B. If the bump is sufficiently large and wide, it can cause an incorrect logic
value to be propagated to the next gate in the path containing net B.

113
Figure 3. Effects of Crosstalk at Different Arrival Times.
A net that receives undesirable cross-coupling effects from a nearby net is called a victim net. A
net that causes these effects in a victim net is called an aggressor net. Note that an aggressor net
can itself be a victim net; and a victim net can also be an aggressor net. The terms aggressor and
victim refer to the relationship between two nets being analyzed. The timing impact of an
aggressor net on a victim net depends on several factors:
• The amount of cross-coupled capacitance
• The relative times and slew rates of the signal transitions
• The switching directions (rising, falling)
• The combination of effects from multiple aggressor nets on a single victim net
Figure 3 illustrates the importance of timing considerations for calculating crosstalk effects. The
aggressor signal A has a range of possible arrival times, from early to late. If the transition on A
occurs at about the same time as the transition on B, it could cause the transition on B to occur
later as shown in the figure, possibly contributing to a setup violation, or it could cause the
transition to occur earlier, possibly contributing to a hold violation. If the transition on A occurs
at an early time, it induces an upward bump or glitch on net B before the transition on B, which
has no effect on the timing of signal B. However, a sufficiently large bump can cause
unintended current flow by forward-biasing a pass transistor. Similarly, if the transition on A
occurs at a late time, it induces a bump on B after the transition on B, also with no effect on the
timing of signal B. However, a sufficiently large bump can cause a change in the logic value of
the net, which can be propagated down the timing path. In [7], they suggested the codes and
encoder and decoder theoretically. We approached in different way with different Codes words
and implemented using the Synopsys Design Compiler and Design Analyzer and provided the
timing results.

114
3. PROPOSED CODES FOR ADDRESS BUSES
Table 1. Codes for Address Buses.
Brute-Keutzer codes Our codes
000 0111 000 0000
001 0001 001 0010
010 1111 010 1010
011 0000 011 1110
100 0101 100 1000
101 0100 101 0011
110 1101 110 1011
111 1100 111 1111
The above Table 1 shows the codes proposed by Brute-Keutzer theoretically and codes
proposed by us. All the three bit codes are encoded to four bit code words which eliminates the
crosstalk. If there are 24 lines, then we have to incorporate an extra 8 lines to eliminate the
crosstalk. All the codes are designed and simulated, synthesized and verified using Synopsys
Tools.
4. GATE LEVEL ENCODER AND DECODER
Figure 4. 3x4 encoder gate level schematic.
Figure 5. 4x3 decoder gate level schematic.

115
Figure 4 and 5 shows the gate level diagram of an encoder and decoder that could be
implementing on System on chip buses. It is easy to recognize the encoder as a simple
combinational circuit which occupies very less area and the decoder is a function only of the
current code word with no feedback at all. This encoders and decoders can eliminate crosstalk
due to coupling capacitances.
5. EXPERIMENTAL RESULTS
Table 2. Timing Reports for Encoder.
Point Incr. Path
input external delay
0.00 0.00 f
x[1] (in) 0.00 0.00 f
U14/Z (NR2) 1.29 1.29 r
U12/Z (NR2) 0.28 1.57 f
y[0] (out) 0.00 1.57 f
data arrival time 1.57
The above timing report shows the data arrival time for the encoder from input side to output
side. Prime Time reports the worst delay path from input to output. The x [1] is the input port of
the encoder and the name in the bracket is the reference name for that port. Incr is the
incremental path delay. The delay from input port, x[1] to the output of nor gate, (U14/Z) is
1.29 ns. The delay from the output of nor gate, (U14/Z) to the output of next nor gate, (U12/Z)
is 0.28 ns. The output of nor gate, (U12/Z) is the output port of encoder, y(0). So, the total delay
from input port, x[1] to output port y[0] is 1.5 ns. The f in the third column indicates a transition
from 0→1 and r indicates a 1→0 transition.
Table 3. Timing Reports for Decoder.
Point Incr. Path
input external delay
0.00 0.00 r
y[1] (in) 0.00 0.00 r
U12/Z (IV) 0.32 0.32f
U8/Z (AO6) 1.22 1.54 r
U7/Z (IV) 0.18 1.72f
x[2] (out) 0.00 1.72 f
data arrival time 1.72
The above timing report shows the data arrival time for the decoder from input side to output
side. Prime Time reports the worst delay path from input to output. y[1] is the input port of the
decoder and the name in the bracket is the reference name for that port. Incr is the incremental
path delay. The delay from input port, y[1] to the output of not gate, (U12/Z) is 0.32 ns. The
delay from the output of not gate, (U12/Z) to the output of next and-or-invert gate, (U8/Z) is
1.22 ns. Next, the delay from the output of and-or-invert gate, (U8/Z) to the output of next not
gate, (U7/Z) is 0.18 ns. The output of not gate, (U7/Z) is the output port of decoder, x(2). So,

116
the total delay from input port, y[1] to output port x[2] is 1.72 ns. The f in the third column
indicates a transition from 0→1 and r indicates a 1→0 transition.
6. COMPARISON TO OTHER TECHNIQUES
This paper is based on the Theoretical concepts proposed in [7]. The encoder presented in fig 4
that takes the code words of our codes has a data arrival time of 1.5ns compared to encoder
proposed in [7] has a data arrival time of 1.75ns. When the data arrival time is small, that
creates a more positive slack. In the literature, there are a number of other techniques designed
for combating crosstalk. Many of them, such as those described in [1], [2] and [3], employ
creative routing strategies in order to minimize crosstalk delay within a data path or logic block.
Our technique on the other hand, is intended for use with long, straight buses, and thus these
routing schemes are not applicable to our domain of interest. [8] and [9] mention some
techniques that are more relevant, such as skewing the timing of signals on adjacent wires,
interleaving mutually exclusive buses, and precharging the bus. However, skewing requires
careful, technology-dependent circuit design and brings up tricky timing issues, whereas our
technique is technology-independent and fully synchronous, with the crosstalk immunity
“correct by construction.” Interleaving is a useful technique, but it cannot be used with buses
that are allowed to transition on any and every clock cycle. Precharging a long bus can incur
detrimental power costs, and is usually not an option. Probably the most common technique is
simply using large repeaters to drive the Miller capacitance through brute force [10]. A
quantitative comparison between our technique and optimally-sized repeaters is technology- and
implementation-dependent. However, conceptually, using large repeaters is a power-hungry
technique, and shielding is an area-hungry technique. Crosstalk-immune bus encoding avoids
crosstalk delay with a modest impact on either area or power.
7. CONCLUSIONS
Cross-talk between wires of an on-chip bus is becoming a significant problem in deep sub-
micron IC design. Crosstalk can result in significant delay variations as well as signal integrity
problems. In this paper, we have introduced the concept of using data encoding to mitigate
crosstalk delay on buses, and we presented a practical framework for understanding crosstalk
immune coding. Current research is to design an efficient encoder and decoder to reduce
crosstalk in on chip buses. The encoder and decoder proposed in our paper reduce crosstalk
delay about 14% to that of available techniques.
ACKNOWLEDGEMENTS
The authors would like to thank Synopsys Tools providers. Without their support, research
would have been a lot more painful experience than it already is.
REFERENCES
[1] Hui Liu, Hong Shi, John Xie “Reduce Simultaneous Switching Jitter in Number, Spatial, Time,
and Frequency Dimensions” IEEE Conf. on Electronic Components and Technology, page no.
1468–1473, 2010.
[2] Chaeho Chung, Soobum Lee, Byung Man Kwak, Gawon Kim, and Joungho Kim “Delay Line
Circuit Design for Crosstalk Minimization Using Genetic Algorithm” IEEE Trans. on Computer-
Aided Design of Integrated Circuits and Systems, VOL. 27, NO. 3 March 2008.

117
[3] Chunjie Duan; Gulat. K.; Khatri S.P.; “Memory-based crosstalk canceling CODECs for on-
chip buses” IEEE Int. Symp. on circuits and systems. pp. 1119-1122, 2006.
[4] P.Heydari, “Capacitance Coupling Noise in High Speed VLSI Circuits,” IEEE Trans. Computer
aided design of integrated circuits, vol.24, march. 2005, page no. 478-489.
[5] P. Subrahmanya and R. Manimegalai and Muthyam, Madhu and V, Kamakoti “A Bus Encoding
Technique for Power and Crosstalk minimization” IEEE International Conf. on VLSI Design,
page no. 443 – 448, 2004
[6] Xiaoliang Bai and Sujit Dey “High level Crosstalk defect simulation methodology for System on
Chip Interconnects” IEEE Trans. on Computer-Aided Design of Integrated Circuits and Systems,
VOL. 23, NO. 9 September 2004.
[7] Bret Victor and and Kurt Keutzer “Bus Encoding to Prevent Crosstalk Delay” IEEE/ACM Int.
Conf. page no. 57-63, 2001.
[8] T. Xue, E. Kuh, and D. Wang, Tost Global Routing Crosstalk Synthesis,” IEEE Trans.
Computer-Aided Design, vol. 16, no. 12.
[9] J. Yim and C. Kyung, “Reducing Cross-Coupling among Interconnect Wires in Deep-Submicron
Datapath Design,” Proceedings. 1999 Design Automation Conf., pp. 485-90. 1999
[10] K. Hirose and H. Yasuura, “A Bus Delay Reduction Technique Considering Crosstalk,” Proc.
Design, Automation and Test in Europe Conf. and Exhibition 2000, pp. 441-5.2000.
[11] D. Li, A. Pua, P. Srivastava, and U. KO, “A Repeater Optimization Methodology for Deep Sub-
Micron, High-Performance Processors,” Proc. Int. Conf. on Computer Design, VU1 in
Computers and Processors. pp. 726-31. 1997
[12] S.P. Khatri, A. Mehrotra, R.K. Brayton, Ralf H.J.M. Otten, and A.L. Sangiovanni-Vincentelli. A
novel vlsi layout fabric for deep sub-micron applications. In Proceedings of Design Automation
Conference, pages 491–496. IEEE, 1999.
[13] C. Duan and S.P. Khatri. Exploiting crosstalk to speed up on-chip buses. In Proceedings of
Design Automation, and Test in Europe (DATE) Conference, 2004.
[14] M. Ghoneima and Y. Ismail. Optimum positioning of interleaved repeaters in bidirectional
buses. Transactions on Computer-Aided Design, 24(3):461–469, 2005.
[15] CN. Satyanarayana, M. Mutyam, and A.V. Babu. Exploiting On-Chip Data Behavior for Delay
Minimization. In Proceedings of International Workshop on System-Level Interconnect
Prediction, Austin, Texas, USA, 2007. IEEE/ACM.

118
Authors
J.Venkateswara Rao received B.Tech Degree from, VRSEC, Nagarjuna
University, Guntur, in 2000. M.Tech. Degree from NITW, Warangal, India in
2003, pursuing Ph.D in the Department of Electronics and Communication
Engineering, JNT University, Hyderabad, INDIA. Currently he is working as
Associate Professor in the Department of Electronics and communication
Engineering, VITS, Hyderabad, INDIA, His research interest includes on and off
chip crosstalk noise reduction in VLSI Circuits.
Dr. Alapati V.Naga Ratna Tilak received his Ph.D. (Micro Electronics) from IIT
Madras. Currently he is working as Prof. in ECE & Dean, Academic Affairs in
Gudlavalleru Engineering College, Gudlavalleru, Krishna Dist. Andhra Pradesh.
India. Currently he is guiding 4 Ph.Ds under JNT University. He is life Member of
IEEE.

A BUS ENCODING TO REDUCE CROSSTALK NOISE EFFECT IN SYSTEM ON CHIP

More Related Content

What's hot (17)

Similar to A BUS ENCODING TO REDUCE CROSSTALK NOISE EFFECT IN SYSTEM ON CHIP (20)

Recently uploaded (20)

A BUS ENCODING TO REDUCE CROSSTALK NOISE EFFECT IN SYSTEM ON CHIP