Dr. Sanjay Banerjee is
the Cockrell Family
Regents Chair Professor
of Electrical and
Computer Engineering and
the Director of the
Microelectronics
Research Center in the
Department of Electrical
& Computer Engineering
at The University of
Texas at Austin. He
received his B.Tech from
the Indian Institute of
Technology, Kharagpur,
and his M.S. and Ph.D.
from the University of
Illinois at
Urbana-Champaign in
1979, 1981 and 1983
respectively, all in
electrical engineering.
As a Member of the
Technical Staff,
Corporate Research,
Development and
Engineering of Texas
Instruments Incorporated
from 1983-1987, he
worked on polysilicon
transistors and dynamic
random access trench
memory cells used by
Texas Instruments in the
world's first 4Megabit
DRAM, for which he was
co-recipient of the Best
Paper Award, IEEE
International Solid
State Circuits
Conference, 1986.
Novel Low Power Transistors in 2D Dirac Materials:
Graphene and Topological Insulators
The semiconductor industry is
placing increasing emphasis on emerging materials
and devices that may provide solutions to
end-of-the-CMOS-roadmap problems beyond 2020. The
remarkable electronic properties of graphene and
topological insulators can lead to novel beyond-CMOS
logic and memory device concepts. One such concept,
the Bilayer pseudoSpin FET (BiSFET), could
potentially be an ultra-low power replacement for
the MOSFET. The BiSFET is based on many-body
tunneling of an excitonic Bose condensate in double
layer graphene. One can also have single particle
2D-to-2D tunneling FETs based on the BiSFET
architecture, with similar current-voltage behavior
demonstrating negative differential resistance.
Graphene has a single-atom
thickness, zero bandgap, symmetric conduction and
valence band structure around the Dirac point, and
low band edge density of states, which combined
makes it theoretically attractive for the formation
of an excitonic condensate. SPICE simulations
demonstrate that fundamental Boolean gates such as
inverters, OR and NAND gates can be implemented with
the BiSFET, using a 4-phase clocked power supply,
with much lower power dissipation and comparable or
faster switching than CMOS. In addition, more
complicated circuits such as 4-bit adders, as well
as simple memory elements, can be achieved leading
to the possibility of both sequential and
combinatorial logic. Experimentally, the BiSFET has
not been demonstrated so far. Thinner (~1nm)
interlayer and lower-k dielectrics than currently
used appear to be required and are now being
explored.
The basic BiSFET device
structure could also be used as 2D-2D single
particle tunnel FETs. For the single particle h-h
and e-e 2D-2D tunnel FETs, graphene’s single-atom
thickness could lead to more ideal interlayer
tunneling characteristics, provided the layers can
be aligned. Single particle tunneling current
calculations have been performed which show NDR
characteristics, reminiscent of the BiSFET, albeit
with higher operating powers.
Finally, the spin-helically
locked Dirac cone surface states in a topological
insulator can be the basis of new types of
non-volatile memory and logic. Several such ideas
will be briefly discussed.
Phil Emma
Chief Scientist, IBM Research
IBM Thomas J. Watson Research Center, NY
Dr. Phil Emma, Chief Scientist at IBM Research, got his
PhD in Electrical Engineering from the University of Illinois a while ago,
and joined the IBM Watson Research Laboratory to work for von Neumann’s
Chief Engineer. He has worked in the areas of microarchitecture,
architecture, systems design, circuit design, electronic packaging,
interconnection technology, and optics. He has also served as an adjunct
professor at several universities. He has written well over 100 technical
articles and several book chapters, and holds over 150 patents. He has
worked as a Design Lead in Development, and led the Systems Technology and
Microarchitecture group at IBM Watson, which works on exploring and
leveraging new physical technologies within the context of systems design.
He has spent the last few years looking at opportunities for 3D in
Architecture. He is a member of the IBM Academy of Technology, and a Fellow
of the IEEE.
Some Future Dimensions
in Scaling
As we all know, we are (again)
approaching the end of CMOS scaling, but this time
it’s true. As (primarily) an Architect, I think that
what the CMOS roadmap did for us was to make
substantial innovation unnecessary. At a high level,
the computers that we make today are (operationally)
the same as the one that von Neumann made ages ago,
albeit with some whistles and bells. I think that
today, we have an opportunity and a motivation to
take what we call “computing” into some new
dimensions using new technology innovation, some new
circuits, and some new concepts of “state.” While
there are many directions, I will discuss
(primarily) 3D technology, and touch on a few more,
with the caveat that today, lots of this will be
consumer-driven.
Rajesh K. Gupta
Professor and Qualcomm Endowed Chair
Department of Computer Science and Engineering
University of California, San Diego
http://mesl.ucsd.edu/gupta/
Rajesh Gupta is a professor and holder of the
QUALCOMM endowed chair in Embedded Microsystems in
the Department of Computer Science & Engineering at
UC San Diego, California. He leads the
Microelectronic
Embedded Systems Lab and is head of the
Embedded Systems
Group at UCSD. Rajesh did his undergraduate
education at IIT-Kanpur and his graduate education
at UC Berkeley and Stanford. He currently serves as
an advisor to Tallwood Venture Capital, RealIntent,
Calypto and Packet Digital Corporation.
Variability-Resistant
HW/SW Stack Through Improved Sensing
Variability in delivered
performance by sensing and computing devices is a
growing reliability challenge. Manufacturing
variability has traditionally been handled through
overdesign by hardware system designers.
Unfortunately, the overdesign margins have become
prohibitively expensive and infeasible with the
growing scale of designs. In an alternate universe,
sensing of the physical environment could provide
important data to adjust software/computation at
different levels. In this talk, I will discuss our
experiments to characterize variability, program
structuring and task scheduling that make a software
stack robust against variations in the computing
environment.
This talk represents part of
the effort at NSF Expeditions in Computing program
on Variability.
Dr. Ian A. Young
Senior Fellow and
Director of Exploratory Integrated Circuits,
Components Research
Intel Corporation, Hillsboro, Oregon
Ian Young is a
Senior Fellow and director of Exploratory Integrated
Circuits in the Technology and Manufacturing Group
of Intel Corporation. He joined Intel in 1983 and
his technical contributions have been in the design
of DRAMs, SRAMs, microprocessor circuit design,
Phase Locked Loops and microprocessor clocking,
mixed-signal circuits for microprocessor high speed
I/O links, RF CMOS circuits for wireless
transceivers, and research for chip to chip optical
I/O. He has also contributed to the definition and
development of Intel’s process technologies.
He now leads a research group exploring the future
options for the integrated circuit in the beyond
CMOS era. Recent work has developed a uniform
benchmarking to identify the technology options in
spintronics,
tunneling junction
field-effect and
photonics
devices.
Ian Young received the Bachelor of Electrical
Engineering and the Master Eng. Science, Microwave
Communications, from the University of Melbourne,
Australia. He received the PhD in Electrical
Engineering from the University of California,
Berkeley. He is the recipient of the 2009
International Solid-State Circuits Conference's Jack
Raper Award for Outstanding Technology Directions
paper. He is a Fellow of the IEEE.
Exploring the Beyond CMOS Options
for Energy Efficient Computation
CMOS integrated
circuit technology for computation is at an
inflexion point. Although this is the technology
which has enabled the semiconductor industry to make
vast progress over the past 30-plus years, it is
expected to see challenges going beyond the ten year
horizon, particularly from an energy efficiency
point of view. Thus it is extremely important for
the semiconductor industry to discover a new
integrated circuit technology which can carry us to
the beyond CMOS era, so that the power-performance
of computing can continue to improve. Currently,
researchers are exploring novel device concepts and
new information tokens as an alternative for CMOS
technology. Examples of areas being actively
researched are; quantum electronic devices, such as
the tunneling field-effect transistor (TFET),
devices based on electron spin and nano-magnetics (spintronics),
and synchronization of coupled- oscillator arrays
for non-boolean logic associated memory. It is clear
that choices will need to be made in the next few
years to identify viable alternatives for CMOS by
2020. To prioritize and guide the research
exploration, benchmarking methodology and metrics
are being used.
This talk will give an overview of the beyond CMOS
device research horizon and the benchmarking of
these devices for computation. A more detailed
investigation of circuits based upon some promising
beyond-CMOS devices will follow. First the
differences of TFET circuits from CMOS will be
described together with their potential for
significant improvement in computational energy
efficiency. Next spintronic devices will be
discussed. These are potential candidates to
complement CMOS technology due to the promise of
non-volatility and compact implementation of logic
function. One promising device, called the All-Spin
Logic (ASL) device will be explained. A complete
static ASL logic family comprised of majority logic
gates, operating from 10mV power supplies, has been
functionally validated with physics-based models and
a spin circuit theory based circuit simulator. Along
with the implementations of sequencing elements
(e.g., latch and D flip-flop), clocked ASL can
implement all the functional blocks for general
purpose computation. The last topic introduces a
novel concept to improve computational energy
efficiency for pattern recognition. It will describe
a non-boolean logic associative memory based upon
arrays of synchronizing coupled oscillators.