The Didactical
Value
of Moore’s Law in Computer and Information
Technology
George Teston, Ph.D.
Central Michigan University
Moore’s Law, a postulate of Intel co-founder Gordon Moore,
states that
the number of transistors on a chip doubles every 18 to 24 months
while transistor
size actually shrinks (Intel, 2002). Moore first hypothesized the
relationship
between circuit density and development time in Electronics
magazine
when in 1965 he wrote:
The complexity for minimum component cost has increased at a
rate
of roughly a factor of two per year. Certainly over the short term
this rate
can be expected to continue, if not to increase. Over the longer term,
the rate
of increase is a bit more uncertain, although there is no reason to
believe
it will not remain nearly constant for at least ten years. (Moore,
1965).
Officially Moore’s Law states that circuit density or capacity
of semiconductors
doubles every eighteen months or quadruples every three years. In
algorithmic
form this is represented as: Circuits per chip =
2(year-1975)/1.5
The popular press of the information technology industry has referred
to Moore’s
Law as "a cornerstone of our culture and economy, something akin
to a constitution
for the information age" (Fixmer, p. 40). Others have touted it
as "the
very principle that governs the information age" (Malone, p. 53).
Not only
has Moore’s prediction been validated, it serves as a modern
benchmark
for the semiconductor industry to maintain a rate of exponential
growth. According
to Schaller (1996), "Moore’s Law produces organizing and
coordinating
effects throughout the semiconductor industry that not only set the
pace of
innovation, but define the rules and very nature of competition"
(p. 1).
He adds, "…the impact of Moore’s Law has led users and
consumers
to expect a continuous stream of faster, better, and cheaper
high-technology
products." Indeed, the economic, technological, and industrial
implications
of Moore’s Law are considerable.
Given its current and future gravity, it is the position of this
researcher
that Moore’s Law should have a greater presence in computer and
information
technology curricula. While Moore’s Law is briefly mentioned in
most computer
textbooks, no apparent research or body of educational discourse
appears to
explore its pedagogical benefits. This paper attempts to describe the
past and
future of Moore’s Law along with its broad implications. This
paper further
hypothesizes the benefits of related discussion and activities within
the college
classroom for computer science and information technology
students.
2. Background
2.1 Historical Perspectives
Moore first made the empirical assertion based on three cycles of
integrated
circuit innovation–"from a single transistor in 1959 to 32
transistors
on a chip in 1964 to 64 transistors in 1965" (Fixmer, p. 40). Not
only
did Moore’s prediction hold true into the next decade, it proved
germane
over the next four decades. The 64-transistor chip that
provided the
crest of his 1965 baseline data grew to a chip boasting 42 million
transistors
by 2000, the Pentium 4.
Figure A, Dickson and DeSanctis (2001)
2.2 Research and Development
To keep pace with Moore’s Law, Intel increased its R&D
spending to
its highest level ever, an estimated $4.1 billion in 2002 (S.B.,
2002). Intel
is opening four new identical fabrication plants by 2004 at a cost of
more than
$10 billion, a controversial move during lean economic times for the
technology
sector (Schlender, 2002). Intel anticipates unprecedented performance
in its
64-bit architecture chip, which it plans to release in 2004. Code
named "Montecito,"
the chip will utilize stretched silicon technology to speed the flow
of electrons
and reduce size. With transistors measuring 50 nanometers, Montecito
will accommodate
a significantly higher number of transistors than the current 64-bit
chip, the
Itanium 2 (McDougall, 2002). According to Schlender (2002), ten of
these transistors
would fit in the diameter of a human hair. This announcement confirms
that Intel
already has production-stage technology congruent with Moore’s
Law through
2004. Intel has also demonstrated that it has laboratory-stage
technology with
a functional silicon transistor of only 15 nanometers, resulting in a
10 GHz
chip that is over four times faster than today’s Pentium 4.
Industry leaders
doubt that even such technology would sustain Moore’s Law beyond
the year
2015. According to Fixmer (2002), "…fabrication will soon
reach limits
at which thresholds only a few atoms thick, probably just under 9
nanometers,
fail to control the flow of electrons, thus negating the
semiconducting properties
of silicon" (p. 39). Current Intel CEO Craig Barrett refutes such
predictions,
adding "Moore’s Law is good for another 15 years"
(S.B., 2002).
Leading chip manufacturers have begun using Extreme Ultraviolet
Lithography
(EUV), a technology that Intel says, "will allow us to keep on
the Moore’s
Law path with a new technology generation every two years"
(Spooner, 2002).
Photolithography, the standard manufacturing technology, shrinks and
prints
an image of a circuit pattern on a silicon wafer. The desired pattern
of extremely
small copper circuits is then created by carefully etching away layers
of metal.
The EUV technology utilizes a significantly higher wavelength to
capture the
image of the circuit grid. The smaller wavelength makes smaller
circuits possible,
thereby increasing the potential transistor density of the silicon
chip.
For the immediate future, Intel already has plans to release new
microprocessors
that it hopes will maintain Moore’s Law and surpass its main
competitor,
AMD. Specifically it plans to release a 2.8 GHz Pentium 4 in the
desktop market
in the second fiscal quarter of 2003. By the end of the forth quarter
it plans
to release a 3 GHz Pentium 4 and a 2.2 GHz Pentium 4 for notebooks
(Kanellos,
2002).
2.3 Future Technologies
As the limits of silicon technology are approached, innovations in
three-dimensional
circuit design will be the next avenue to uninterrupted chip growth.
"New
paradigms such as molecular transistors, carbon nanotube gates and
quantum computing
will probably continue for many decades to produce growth at roughly
the same
or an even greater exponential rate than the curve described by
Moore’s
Law" (Fixmer, p. 39). Inventor and well-noted technology author
Ray Kurzweil
explores these prospects in his new book The Singularity In
Near (2002).
He believes that silicon will give way to the nanotube, which will
give way
to quantum computing, thereby surpassing Moore’s Law at a
double
exponential rate.
Within the past year, the research community has made major strides
in creating
three-dimensional carbon nanotube circuits. These new logic gates are
constructed
using a chemical process to create a tubular carbon molecule rolled
from a graphite
sheet one atom thick. The result is a circuit many times smaller and
stronger
than conventional silicon.
In October of 2001, the Bell Labs division of Lucent Technologies
successfully
created functional organic transistors at a molecular level. According
to Fixmer
(2002), molecular transistors would enable "density increases
within the
next 15 years of around 106 times today’s most
advanced silicon
chips–a threshold of computing power that could support speech,
sensory
and decision-making function approximating human intelligence"
(p. 40).
Quantum computing, though still largely a theoretical pursuit, holds
promise
beyond nanotube technology. The position of an electron within a
single atom
could be manipulated to created a range of logic states, called
qu-bits. "Unlike
a bit, which must be 1 or 0, a qu-bit remains in a simultaneous state
of both
0 and 1 until an event forces it to decide" (Fixmer, p. 40).
2.4 The Impact of Moore’s Law
Moore’s Law is widely considered to be an important barometer of
our technological
evolution, and its longevity has spawned a great deal of debate.
Critics argue
that Moore’s Law is approaching its limit as the laws of physics
become
a barrier to silicon etching. Meanwhile, Intel races to continue the
exponential
growth rate of transistor density within the silicon paradigm.
Regardless of
how Moore’s Law fits within the context of future technologies,
its importance
today cannot be overstated. Wall Street accepts Moore’s Law as
"a
prescription for uninterrupted market growth and steady improvement in
worker
productivity" (p. 39). When advances or failures are made within
the semiconductor
industry, an economic ripple effect reaches across the technology
sector.
In addition to Intel, Moore’s Law is important to Advanced Micro
Devices,
IBM, Infineon, Micron Technologies, Motorola, and throughout the
semiconductor
industry. "Moore’s Law is important because it is the only
stable
ruler we have today, … it’s a sort of technological
barameter. It
very clearly tells you that if you take the information processing
power you
have today and multiply it by two, that will be what your competition
will be
doing 18 months from now. And that is where you too will have to
be." (Malone,
1996). Barret, Intel CEO, states that, "It’s a fundamental
expectation
that everybody at Intel buys into" (Schlender, 2002, p. 98).
Moore’s Law also has direct implications for the software
industry. Microsoft’s
Advanced Technology Group conducted a survey to gauge the growth of
software
processing requirements. The researchers measured a variety of
Microsoft products
by counting the lines of code used in successive releases of the same
title.
Basic had 4,000 lines of code in 1975; 20 years later it had
almost half
a million. Microsoft Word had 27,000 lines of code in the
initial release
in 1982; by 1996 version 6.0 had over 2 million (Schaller, 1996).
Thus, a relationship
can be drawn between Moore’s Law and software. The size and
complexity
of software has grown even faster than Moore’s Law, supporting a
market
for faster processors (Schaller, 1996). Because virtually all
electronics today
utilize semiconductors, Moore’s Law bears directly on the future
of a multitude
of devices and the industries that support them. Children’s toys,
cell
phones, satellites, weapon systems, and corporate Web servers all rely
on semiconductor
innovation and therefore will likely advance relative to Moore’s
Law. Schlender
(2002) hypothesizes the direct impact of Moore’s law on the
future telecom
sector by stating, "The chips allow notebooks to speak wirelessly
to networks,
enable cell phones to make calls, and help route Web pages, e-mail,
and stream
media around the Internet" (p. 100).
The impact of Gordon Moore’s premise reaches all the way to
Congress,
where in 2000 the Joint Economic Committee reported, "The
efficiency gains
in I.T. production, particularly semiconductors, will eventually run
into physical
constraints; Moore’s Law cannot hold indefinitely" (Fixmer,
p. 40).
If exponential growth of microprocessors is considered so fundamental
to our
society, economy, and future industrial outlook, surely any modern
computer
curriculum should address Moore’s Law. The principles of this
40-year old
prediction provide a perfect framework to understand the evolution of
our computing
ability, to see from where we came and to postulate about where we can
go. Ironically,
a search of ERIC, the leading educational research database, yielded
no articles
related to Moore’s Law within computer science education or
information
technology education. The same search on Google, the leading Internet
meta-search
engine, yielded nothing to support a curricular nexus of this
important concept.
3. Methodology
The sample used for this study is comprised of 72 undergraduate
students from
four introductory computer information technology courses taught over
a period
of one year by the same professor. The setting is a university in
Atlanta, Georgia.
The convenience sample used in this study represents "action
research"
and therefore does not enjoy the same validity and reliability as a
formal research
design. Data and observations are from the regular instructional
activities
of the course and are not the product of tightly controlled
variables.
To minimize the possibility of a Hawthorne effect from one term to
the next,
no special instrument related to Moore’s Law was used nor were
the students
told that observations were being made with regard to Moore’s
Law. Instead,
items related to Moore’s Law from routine class assessments were
used along
with a class graphing activity. The same lecture notes and PowerPoint
presentations
were used from term to term to ensure consistency of delivery of the
Moore’s
Law concepts.
Before being exposed to the concept of Moore’s Law, students
were taught
about different microprocessor generations, the transistor density,
and speeds
of each. The professor provided explanation about the relationship
between transistor
density and speed performance. Students in classes 1 (fall term) and 3
(spring
term) were then asked to plot each of the chips on a graph in relation
to the
time of release (X axis) and the transistor density (Y axis) for each
chip.
Students in classes 2 (winter term) and 4 (summer term) were provided
a pre-made
graph of the same chip density to time relationships. (See Figure
C.)
Using the graphs, students were asked to comment about the historical
rate
of chip capacity growth, predict future speeds, and estimate a
specific growth
factor. After the results of the graphing activity were collected, the
professor
explained the concept of Moore’s Law. Students’
comprehension of the
Moore’s Law concept was then tested at the end of the computer
architecture
unit by two objective, multiple choice questions about Moore’s
law and
two open-ended questions that asked:
A. Based on your observation of microchip innovation over
the last
40 years, how quickly are new chip capacities being doubled?
B. Based on that rate of innovation, what type of change do you
predict for
microchip-based devices in the future?
4. Findings
For classes 1 and 3, the group (n=33) allowed to freely create their
own scale
and graph of the transistor capacity points, 19 students (57.7%)
created a proportional
Y-axis scale for transistor capacity. The other 14 students (42.4%)
created
a logarithmic scale for transistor capacity. (See Figure B.) Students
who attempted
to create the graph with a proportional scale appeared to have
difficulties
allowing an adequate range to plot the points with any substantial
degree of
discrimination.
The proportional scale students demonstrated an excellent
understanding of
the historical rate of microprocessor innovation, with 85.7%
indicating the
rate to be "exponential" on the objective portion of the
assessment.
When asked to estimate the rate at which the capacity doubles in
open-ended
growth question A, responses ranged widely from 3 months to 4 years.
This same
group described mainly lifestyle changes in the open-ended prediction
question
B.
The logarithmic scale students demonstrated a relatively weak
understanding
of the historical growth rates from interpretation of their graphs.
(See Figure
B.) Only 25.5% indicating the rate to be "exponential."
Surprisingly,
students’ mathematical logic was not congruent between their
ability to
create a logarithmic scale and then interpret the exponential
character of the
points on the graph. When asked to estimate the rate at which the
capacity doubles
in open-ended growth question A, responses ranged from 24 months to 60
months.
This group described mainly different types of devices and increased
device
speeds in the open-ended prediction question B.
Figure B
For classes 2 and 4, the group (n=39) given the pre-made graph from
the Intel
Web site (see Figure C), results were largely similar to the
logarithmic scale
created by students in classes 1 and 3. This was not surprising given
that the
graph Intel provides on its Web site is scaled logarithmically.
Figure C
Only 9 students (23.0%) indicated the rate of innovation to be
exponential.
The response range for growth question A was 30 months to 63 months.
Just as
with the logarithmic students from classes 1 and 3, the most common
response
for this question was 60 months (f =14). This was possibly due
to the
fact that the X-axis for time was set into increments of 5 years, or
60 months.
Students likely perceived the slope of the graph to double at roughly
each 5-year
interval on the X-axis. This was consistent with the other logarithmic
group’s
inability to transfer any apparent exponential view of the graph
despite having
created a logarithmic scale. In open-ended prediction question B, this
group
described predominantly future devices characterized by greatly
reduced size
and increased speeds. Only five students (13.8%) referred to any
change in lifestyle
or society resulting from the growth trend depicted in the graph.
5. Summary and Conclusions
Students appeared to have difficulties in creating proportional
graphs of Moore’s
Law because of the enormous range of transistor growth. Exact accuracy
was not
the objective of plotting the chips, but rather the purpose was to
gain a general
understanding of growth trends. Even with accuracy substantially
impaired by
the size limitation of the page on which the graph was being created,
the students
who created a proportional graph appeared to gain a solid grasp of the
historical
perspective of microprocessor innovation. Most of the these students
even recognized
the exponential character of the growth. Yet these students were
largely unable
to make good predictions about the rate at which the transistor
capacities doubled.
This appeared to be caused by the poor degree of detail possible with
a proportional
scale on a graph limited to an 8.5 x 11 page. On the other hand, the
students
were able to describe sweeping lifestyle and societal changes when
asked to
predict future impact, a possible result of their historical
perspectives of
the exponential growth.
While the logarithmic scale students were weaker in terms of
historical perspective
than the proportional scale students, they were better in estimating
the doubling
rate of transistor capacity. This result appeared to be positively
related to
the greater degree of accuracy possible in the logarithmic graph.
Supporting
the relationship between historical perspective and future impact
predictions
observed in the proportional group, the logarithmic group exhibited a
poor grasp
of historical perspective and far less abstract concepts of future
impact. This
result seems to support the notion that one’s understanding of
past technology
directly influences his or her ability to envision future technology
beyond
the conventional to the abstract.
The observations noted here are the result of action research and
therefore
lack the formal controls that permit the sample to be generalized to a
broader
population. Nonetheless, the results suggest a number of possible
issues worthy
of further research. Students seemed to garner a general understanding
of the
principles of Moore’s Law from the graphing exercises. Use of
logarithmically
scaled graphs such as the one Intel provides on its Web site may be
misleading
to some students. Students may interpret the apparent trend without
recognizing
the exponential character concealed by the scale. Proportionally
scaled graphs
may have limited accuracy for transistor growth during the early years
of the
industry because the points appear so close to each other. Because the
proportional
and logarithmic scaling strategies had converse strengths and
weaknesses, students
would likely benefit from exposure to both. Ironically, a search of
the academic
literature revealed nothing related to Moore’s Law within
computer science
education or information technology education. Although Moore’s
Law may
be a common factoid in most computer texts or courses, the lack of
related academic
attention suggests that its didactical value is being seriously
overlooked.
Moore’s Law graphing activities, such as the ones explored here,
appear
to have substantial value for the instruction of computer
architecture, basic
literacy, and information technology innovation. When graphed
properly, this
40-year-old prediction provides a perfect framework for students to
learn the
evolution of our computing ability, to quantify the growth rate of
computer
innovation, and to postulate about revolutionary future information
technology
applications.
6. References
Dickson, G., & DeSanctis, G. (2001). Information technology
and the
future enterprise. Upper Saddle River, NJ: Prentice Hall.
Fixmer, R. (2002). Law and order. eWeek. April 15, 39-40.
Intel Corporation. Moore’s law. [Online document].
Available: www.intel.com/research/silicon/mooreslaw.htm.
Kanellos, M. (2002). Intel pushes faster for new Pentium 4. CNET
News
[Online document]. Available:
http://news.com.com/2100-1001-945684.html?tag=fd_top.
Malone, M. (1996). Chips triumphant. Forbes ASAP. February 26,
53-82.
McDougall, P. (2002). Intel keeps pace with Moore’s law.
Informationweek.
August 19, 28.
Moore, G. (1965). Cramming more components onto integrated circuits.
Electronics,
38(8), 114-117.
Moursund, D. (1998). Moore’s Law. Learning and leading with
technology,
25(6), 4-5.
S. B. (2002). Upholding Moore’s Law. Newsweek. March 25,
47-48.
Schaller, B. (2002). The origin, nature, and implications of
Moore’s Law.
[Online document]. Available:
http://mason.gmu.edu/~rschalle/moorelaw.html.
Schlender, B. (2002). Intel’s $10 billion gamble.
Fortune. November
11, 90-104.
Spooner, J. (2002). Intel breathes life into Moore’s Law.
CNET News
[online document]. Available:
http://news.com.com/2100-1001-888781.html.
Sutherland, J. (n.d.). Evolution of computer power / cost
[online document].
Available: http://jeffsutherland.org/objwld98/future.html.
George Teston, Ph.D., M.S., M.Ed. is an associate graduate professor
of computer
information systems for Central Michigan University. He has also
taught at Georgia
College & State University, The University of Georgia,
Truett-McConnell College,
and Life University. Prior to entering higher computer education, Dr.
Teston
taught computer programming on the high school level and owned a
computer consulting
business for a number of years. His specialties include programming,
multimedia
development, business applications, computer graphic design, web
development,
and computer science. Dr. Teston's research interests include computer
ethics,
software piracy, and computer education.
Contact:
doctorteston@yahoo.com
http://computer-class.com
Copyright ©
2003, ISTE (International Society for Technology in Education). All
rights reserved.
|
