FPL2000 PROGRAM and SITE details: http://www.fpl.uni-kl.de/FPL/
Biology, Molecular Biology, Molecular Computing, Mircoreactor, Information Technology, Microelectronics, Semiconductors, Integrated Circuits, Computers, Soft Hardware, FPGA, Reconfigurable Computing, selfrewiring chips, MEMS, Evolvable Hardware, Embryonic Hardware, Genetic Algorithms, Network Processors, Wireless Communication,
KAISERSLAUTERN / VILLACH --
"Use Configware for Software" is blared out from the bandwagon.
In his keynote scheduled for FPL2000 Hitachi's corporate senior chief technologist
Dr. Tsugio Makimoto envisions the impact of reconfigurable computing becoming
mainstream microchip application, as he has predicted 14 years ago. With
silicon technology progress facing its limits next decade Makimoto's Law
may well overtrump the Gordon Moore Law, says program chair Reiner Hartenstein.
(Gordon Moore, a founder of intel, predicted in 1965: The number of transistors
on a microchip doubles every 18 to 24 months. Dr. Tsugio Makimoto is Hitachi's
corporate senior chief technologist, IEEE fellow, recipient of the prestigious
Ichimura Award, member of the advisory board of Japan's Nara Institute
of Science and Technology (NAIST), member of the international advisory
panel of the National Science and Technology Board (NSTB) of Singapore,
and member of the board of directors of Chartered Semiconductor Manufacturing,
Singapore.)
Makimoto observed in 1986,
that mainstream microchip application changes every ten years: what's called
"Makimoto's Wave". He predicted, that the "third wave" will bring
reconfigurable hardware into mainstream. "Makimoto's third wave has been
started", says Hartenstein: "the Configware Rush is taking off right now,
portraying the microprocessor as a methusela". (Software as we know
it can run only sequential programs on von-Neumann-type hardware. But software
cannot be used for the spatial programming of reconfigurable hardware,
where a different programming medium is needed: "configware". Software
is based on instruction fetch during run time, whereas configware determines
(kind of) "instruction fetch" before run time: drastically more powerful
complex "instructions" by configuring powerful soft datapaths, where at
run time only data streams are piped through, but no instruction streams.
See [1], or, introduction in [2], [3]. - Already now reconfigurable integrated
circuits like FPGAs are a rapidly growing multi billion dollar market,
making ASICs disappear within a few years. See [4], [5]. Analysts predict
around $50 billion per year by end of this decade.)
FPL2000. Founded 1991
at Oxford, UK, FPL is the eldest international conference on reconfigurable
computing. FPL2000, the 10th international conference on Field-Programmable
Logic and Applications, will be held August 27-30 at Villach, Carinthia,
Austria. Its ernomous growth rate confirms Makimoto's prediction. Doubled
attendance is expected, since, compared to FPL'99 at Glasgow, submissions
have doubled: http://www.fpl.uni-kl.de/FPL/ .
Product Longevity. Reconfigurable
macro module cores will be indispensable ingredients for System on a Chip
(SoC) designs, claims Jan Rabaey, head of the Wireless Research Center
and Professor at UC Berkeley, in his FPL2000 keynote: especially for next
generation cellular wireless communication. Soap Chip for SoC" says Reiner
Hartenstein from Kaiserslautern: System-on-a-programmable Chip". We are
heading toward a microelectronics market revolution, says Tom Kean from
Algotronix (Edinburgh) in his keynote: by reconfigurability to cure shrinking
product life cycles at exploding design cost, to obtain product longevity
through new horizons of flexibility.
Without Reconfigurable "Machines"
there is no way out of the current microchip design crisis, says Reiner
Hartenstein, Professor at Kaiserslautern: since the 55 years old von Neumann
scheme does not support soft datapaths like the KressArray and others,
a new machine paradigm which accepts configware will move up with Makimoto's
3rd wave. Hartenstein envisions an emerging dichotomy of programming where
configware engineering competes with software engineering. Configware will
shake not only the foundations of programming, but even entire computer
science and engineering curricula, says FPL2000 general chair Herbert Gruenbacher
of Carinthia Tech. Reconfigurable machines are a revival of the "fixed
plus variable structure computer" idea, published in 1960 by Gerald Estrin,
now a professor emeritus of University of California, Los Angeles,
Merging with Molecular Biology.
FPL2000 also bridges the gap between the scenes of microelectronic reconfigurability
and evolvable systems, and, of molecular computing - by introducing to
exciting new developments in multi-disciplinary co-operations between Computer
Science, Molecular Biology, and other relevant areas, which shake the traditional
definitions of computer science, and, which are supported by major consortia:
the European Molecular Computing Consortium (EMCC), the US "Consortium
for Biomolecular Computing", and, the Japanese "Molecular Computer Project".
Biology versus Microelectronics.
Reconfigurability is not only the basis of biologically inspired microelectronic
systems, but is going beyond traditional electronic context, where molecular
biology is just another reconfigurable medium: less fast, but much more
flexible, says invited speaker John McCaskill, who expects a rapidly increasing
interchange between molecular biology, nanotechnology, microsystems, electronics
and information technology in the transition from reconfigurable systems
to fully evolvable systems - due to the rapid progress in exploring the
parallels between biological reconfiguration and evolution, and, opportunities
for using reconfiguration to evolve complex nano-, micro- and electronic-scale
devices and computers.
Prof. Dr. John McCaskill
is the head of the BioMIP institute of GMD (German National Research Center)
near Bonn, investigating the principles and potentials for natural design
and programmability in complex biomolecular synthesis systems, like microreactors.
His group also works for the NASA on self-reproducing molecular systems
and Darwinian chemistry, and he also is international project coordinator
and has grants to work on DNA computing (molecular computing) and on pattern
formation of molecular ecosystems in microreactor networks.
Microreactors stem from
shrinking the test tube down to microscopic size, so that thousands of
simultaneous tests take place simultaneously on a microchip, used for drug
discovery, genomics, clinical diagnostics, basic research, and industrial
chemical applications - doing within minutes what otherwise would take
hours or days. Prof. Hartenstein from Kaiserslautern says: "Modern microreactors
are doing for fluids what microprocessors and electronic circuits do for
electrons, and will be soon to drug discovery and clinical diagnostics
what Pentium processors are to the computing world: chemical bits for electronic
bits.
Reconfigurable microreactors
are hard on the heels of reconfigurable microelectronic systems, says John
McCaskill: Flow microreactors have a circuit-like network of channels and
gate-like reactorchambers, where microvalves, controlling the passage of
fluid or charged molecules play the role of transistors. For Reconfigurability
microvalves are "the liquid analog to a transistor". Prof. McCaskill points
out, that also the intermediate level of microfluidic systems is open to
reconfiguration and evolution, ready for linking up the electronic and
molecular processing worlds.