Achievements of
Prof. Reiner Hartenstein and his team*
Work done and awards received:
Keynotes and other invited
presentations:
Refereed papers / presentations:
List of Achievements
The first book (1977) on a HDL
introducing
a methodology for Structured VLSI Design - a bestseller (5000
sold):
- R. Hartenstein:
"Fundamentals of Structured Hardware Design
- A Design Language Approach at Register Level"; North Holland Publ.
Co./American
Elsevier (Elsevier Scientific), Amsterdam / New York, 1977
Earlier KARL-related publications:
- R. Hartenstein:
Towards a Language
for the Description of IC Chips; SIGMICRO Newsletter .vol. 4, no. 4,
1973
- R. Hartenstein: A
Halfbaked
Idea on a set of Register Transfer Primitives; SIGMICRO Newsletter vol.
4, no. 3, 1973
The first
regularly scheduled Mead &
Conway style university course "on the continent" (Europe and
Asia): (summer semester
1980 at CS department, University
of
Kaiserslautern)
--- the first in
the world except USA
(starting end of 1978) and UK (starting
winter
semester 1979/80) -- later version:
- R.
Hartenstein: (Standort Deutschland: Wozu noch Mikrochips?) -
Einführung in die Methoden der Technischen Informatik; ITpress
Verlag Bruchsal / Chicago, 1994
Authoring and
implementing KARL-II,
throughout
the 80ies the most successful HDL* (1st rank before Ella). see List
of licensee sites - List
of 2nd party tools
interfaced to it - quotation
index
- KARL-related
literature Implementation and application of KARL has been
supported as the CVT project and the CVS projects (grants: a total of
about 85 million ECU from 1983 - 1990) by the European Union within the
framework of the ESPRIT programme. (In
the
90ies VHDL and Verilog took over.)
- R. Hartenstein: KARL and ABL; in J.
Mermet (editor): Fundamentals and Standards in Hardware Description
Languages; Kluwer Academic Publishers, 1993
KARL-II the
first implemented HDL
including
"wiring operator" primitives
(shuffle, butterfly and many
others) -
useful for module generator generators, structured VLSI design, later
also used by Vasily Moshnyaga
and Hiroto Yasuura, University of Kyoto, for a module generator:
- V. Moshnyaga, H. Yasuura et al.: A
Data-Path
Modules Design from Algorithmic Representations; Pro. WG10.5 IFIP
Workshop on Synthesis, Generation and Portability of Library Blocks for
ASIC Design, March 1992.
- V. Moshnyaga, H. Yasuura et
al.: A
Language for Designing Module Generators; Proc. Synthesis and
Simulation Meeting and International Interchange (SASIMI'92),
pp.383-392, April 1992.
the first
implemented RT level HDL
including
fully fledged chip planning primitives*
(layer expression, slice expressions, and
abutment expressions
with port matching syntax).
- R. Hartenstein: The use of KARL
for the description of Integrated Circuits; iin: R. Piloty
(editor):
Proc. Intl. Workshop on Computer Hardware Description Languages,
Darmstadt, 1974, ACM Lecture Notes, German Chapter ACM 1975
- R. Hartenstein: Wiring Algebra
for Specification of VLSI Design Problems; SIGDA Newsletter, Aug. 1980
- R. Hartenstein: KARL and ABL; in J.
Mermet
(editor): Fundamentals and Standards in Hardware Description Languages;
Kluwer Academic Publishers, 1993
- G. Girardi, R. Hartenstein:
ABL-Specification -Draft-; CVT-report,
Torino,
Italy
/ Kaiserslautern, 1983
- G. Girardi, R. Hartenstein, U.
Welters: ABLED: a RT level Schematic Editor and Simulator user
Interface; Int`l EUROMICRO Symposium, Brussels, 1985, North Holland
Publishing Co, Amsterdam 1985
*) at at that
time µFP (Mary Sheeran, Ph. D. thesis, submitted Nov.
1983) was the
only other HDL
including a topological
operator to express relative position of operators (but port specs have
not not
included)
Authoring and
Implementation of
KARL-III (also called CVS_BK or CVS Behavioral KARL), a KARL-II extension also supporting the
inclusion of procedural descriptions.
- KARL-III Primer (draft), CVT report, Univ. Kaiserslautern, April
1985
- A.Bonomo, G.Girardi, R. Hartenstein, L.Lavagno, R. Hauck: Syntax
Diagrams of the CVS_BK Language (CVS Behavioral KARL); ESPRIT / CVS
report, CSELT, Torino, Italy / Univ. Kaiserslautern, Germany, Febr. 1987
- A.Bonomo, G.Girardi, R. Hartenstein, L.Lavagno, R.Hauck: Semantic
Specification of CVS_BK Language (CVS Behavioral KARL); ESPRIT / CVS
report, CSELT, Torino, Italy / Univ. Kaiserslautern, Germany, Febr. 1987
The
first
graphical HDL editor implemented, also
including floor planning features. The only
earlier graphical HDL having been implemented at University of
Missouri at
Rolla was merely a HDL-oriented schematics entry system.
- R. Hartenstein: The
use of the
block diagram language ABL; in: R. Piloty (editor): Proceedings of the
Intl. Workshop on Computer Hardware Description Languages, Darmstadt,
1974,
ACM Lecture Notes, German Chapter of the ACM 1975
- G. Girardi, U. Welters et al.: ABLED: a RT level
Schematic
Editor
and Simulator user Interface; Int`l EUROMICRO Symposium, Brussels,
1985,
North Holland Publishing Co, Amsterdam 1985
- A. Bonomo, G. Girardi, A. Leece, L. Magiulli: GENMON: a
specialized ABL editor for design methodology descriptions; 2nd ABL and
KARL User group workshop (ABAKUS workshop), Igls, Austria, 1988
Shifting Functional Design Verification
Toward RT
Level
- W. Nebel et al.: Shifting Functional
Design Verification
Toward RT
Level
by Automatic Register Transfer Net Extraction; IFIP CHDL'87, Amsterdam,
1987
- W. Nebel et al.: Functional Design
Verification by Register
Transfer
Net Extraction from Integrated Circuit Layout Data; IEEE COMPEURO,
Hamburg,
1987
- W. Nebel: -REX- Automatic
Extraction of RT-Level Descriptions from Integrated Circuit Layout
Data; Ph. D. Dissertation, University of Kaiserslautern, 1986 -
The first proposed term rewriting (TR)
approach
for structured VLSI synthesis. All
earlier TR approaches in EDA deserved verification (bottom-up
approach), whereas the one proposed by Kaiserslautern is the first one
which
deserves synthesis (top-down approach)
- R. Hartenstein:
Wiring Algebra
for Specification of VLSI Design Problems; SIGDA Newsletter, Aug. 1980
- elaborated design example proposed
around 1985;
- published 1994 in section
20.4.2. Algebraischer Entwurf eines Multiplizierers; in the book: R.
Hartenstein: (Standort Deutschland: Wozu noch Mikrochips?) -
Einführung in die Methoden der Technischen Informatik; ITpress
Verlag Bruchsal / Chicago, 1994
- Adopted
by the TR scene in 2002:
- M. Ayala-Rincón, R. Maya Neto, R. P. Jacobi, C.
Llanos, R. Hartenstein, Applying ELAN Strategies in Simulating
Processors over Simple Architectures, (ps) (pdf) (ppt). In Proc.
2nd
Int. Workshop on Reduction Strategies in Rewriting and Programming -
WRS'02, pages 127-141, Copenhagen, Denmark, July 21, 2002. In Vol.
70, issue 6 of the Elsevier ENTCS
- M. Ayala-Rincón, R. Hartenstein, R. Maya Neto, R. P.
Jacobi and C. Llanos, Architectural Specification, Exploration
and Simulation Through Rewriting-Logic, (ps). Colombian
Journal of Computation, Vol 3(2):20-34, 2003.
- R. Hartenstein, R. P. Jacobi, M. Ayala-Rincón and C.
Llanos, Using Rewriting-Logic Notation for Functional
Verification in Data-Stream Based Reconfigurable Computing, (ps, ; Pdf). In Forum on Specification and Design
Languages - FDL 03, Frankfurt, Germany, (Sep 23-26, 2003).
Invented and
implemented the
data-stream-based anti machine paradigm (also sometimes called Xputer
machine
paradigm (XMP))
using data counters instead of a program counter: the counterpart of
the instruction-stream-based
"von Neumann" machine.
This
is the first general machine paradigm for
processors or arrays using soft data paths.
Its data counters are co-located with (auto-sequencing) data memory
banks (not co-located
with the data path)
Data-stream-based
computing also coming along with the anti machine paradigm and
related compilation methods have recently been recognized as a general
method to cope
with the memory wall, having been the main reason of stalled progress
in high performance computing during the past 2 decades.
- R. Hartenstein, A.G. Hirschbiel, M.Weber: MOM - a partly
custom-design
architecture
compared to standard hardware, Proc. IEEE COMPEURO, Hamburg 1989, IEEE
Press, 1989
- R. Hartenstein, A.G. Hirschbiel, M.Weber: A Novel Paradigm of
Parallel
Computation
and its Use to Implement Simple High Performance Hardware;
InfoJapan'90-
International Conference memorating the 30th Anniversary of the
Computer
Society of Japan, Tokyo, Japan, 1990
- (see above) Invited reprint in Future
Generation
Computer Systems 7, 1991 / 1992, p. 181-198, North Holland
- R. Hartenstein, A.G. Hirschbiel, M. Riedmüller, K.
Schmidt,
M.Weber: A
Novel
ASIC Design Method based on a Machine Paradigm; IEEE-JSSC - Journal of
Solid State Circuits , Vol. 26, No. 7, pp. 975-989, July 1991
invented and
implemented Auto-sequencing
Data Memory principles using a novel address generator methodology
(GAG: generic address generators) needed for anti machine architectures.
The anti machine based on Auto-sequencing Data Memory has recently
been recognized as a general method to cope with the memory wall, the
main reason of stalled progress in supercomputing during the past 2
decades.
- R. Hartenstein, A. Hirschbiel, M. Riedmüller, K.
Schmidt,
M.Weber (second best paper award): A
Novel
high-performance Machine Paradigm Using Auto-sequencing Data Memory;
1991
Hawaii International Conference on Systems Sciences, Koloa, Kauai,
Hawaii,
1991
- R. Hartenstein, M. Herz, M.
Miranda,
E.
Brockmeyer, F. Catthoor (invited paper): Memory
Organisation
for
Stream-based Reconfigurable Computing; 9th IEEE
International Conference on
Electronics,
Circuits and Systems - ICECS
2002,
September 15-18, 2002, Dubrovnik, Croatia - pdf
- pps-ppt
The first data
context transforms to
optimize storage schemes for anti machines by scan pattern transforms (e. g. rotation,
shearing and others) - avoiding address
computation overhead as well as
avoiding or reducing instruction fetch at run
time
...
- A.
Hirschbiel: A Novel Processor Architecture Based on Auto
Data Sequencing and Low Level Parallelism, Ph. D. Dissertation 1991
- M. Herz:
High Performance Memory Communication Architectures for Coarse-grained
Reconfigurable Computing Systems; Dissertation 2001, Kaiserslautern
University
of Technology - pdf
- R. Hartenstein, M. Herz, M.
Miranda,
E.
Brockmeyer, F. Catthoor (invited paper): Memory
Organisation
for
Stream-based Reconfigurable Computing; 9th IEEE
International Conference on
Electronics,
Circuits and Systems - ICECS
2002,
September 15-18, 2002, Dubrovnik, Croatia - pdf
- pps-ppt
The first general purpose high level programming
language for anti machines (MoPL): data-stream-based (data counter
usage) - instead of instruction-stream-based as known from von Neumann.
- A. Ast, J. Becker, et al.:
Data-procedural Languages
for
FPL-based
Machines; 4th Int. Workshop On Field Programmable Logic And
Applications,
FPL'94, Prague, September 7-10, 1994, Lecture Notes in Computer
Science,
Springer, 1994.
- A. Ast, J. Becker, et al.:
Data-procedural Languages
for
FPL-based
Machines; Universität Kaiserslautern, Fachbereich Informatik,
Interner
Bericht, Nr. 264/95, 1995
The first
automatically partitioning
compiler for
software/configware
co-design
- J. Becker:
A Partitioning Compiler for Computers with Xputer-based Accelerators,
Ph.
D. Dissertation 1997, Kaiserslautern University of Technology - URL
- pdf
- K. Schmidt:
A Program Partitioning, Restructuring, and Mapping Method for Xputers,
Ph. D. Dissertation 1994, Kaiserslautern University of Technology
- available from publisher Shaker
Verlag,
ISBN: 3-8265-0495-X
- J. Becker, et al.: A
Profiling-Driven
Hardware/Software
Partitioning
of High Level Language Specification; IFIP International Workshop on
Logic
and Architecture Synthesis, Grenoble, France, 18. - 19. Dezember 1995
The first
generalization of
the systolic array
(KressArray) implemented
along with synthesis tool (DPSS). Originally systolic arrays could be
used only for algorithms with strictly regular data dependencies,
because their synthesis metholds yielded only arrays with uniform
processing elements, and with only linear pipes. Kaiserslauterns new
DPSS synthesis
method, however, supports all applications also with extremely
irregular data
dependencies, so that also the use of reconfigurable platforms makes
sense. This generalization has been obtained by using simulated
annealing instead of linear projection or algebraic methods for
synthesis.
- R. Kress:
A Fast Reconfigurable ALU for Xputers, Ph. D. Dissertation 1996,
Kaiserslautern
University of Technology - please, request
printed version
- R. Kress et al.: A Data path Synthesis
System for the
Reconfigurable
Data path
Architecture; Asia and South Pacific Design Automation Conference,
ASP-DAC'95,
Nippon Convention Center, Makuhari, Chiba, Japan, Aug./Sept. 1995
The first
design space explorer
implemented
for coarse grain reconfigurable arrays, supporting a wide
variety of array architectures and PE architectures - also including
architectures requiring extremely rich interconnect resources.
- U. Nageldinger:
Coarse-grained Reconfigurable Architectures Design Space Exploration;
Ph. D. Dissertation,
2001 - pdf
- U. Nageldinger et al.:
Design-Space
Exploration
of Low Power Coarse Grained Reconfigurable Data path Array
Architectures;
Proc. PATMOS 2000, International Workshop - Power and Timing
Modeling,
Optimization and Simulation, Göttingen, Germany - September 13-15,
2000
Reiner Hartenstein predicted the decline of
Germany's role as a High
Tech Innovator - already in 1994 and even much earlier.
- R. Hartenstein: Null Bock auf
High Tech; ITpress Verlag, 1994 --
intensions of this book
- R. Hartenstein: Articles in nationwide
newspapers.
Also see here.
Lobbying
efforts in curriculum
development.
Hartenstein has been interested in
Computer architecture education already in the 70ies.
- R. Hartenstein: Increased Hardware Complexity
- A Challenge to Computer Science Education; 1st Int'l. Symposium
for Computer Architecture (ISCA), Gainesville, Florida, Dec. 9. - 11,
1973
More recently Reiner Hartenstein
lobbies for upgrading CS-related curricula to keep future graduates
competitive in the job market in a world of off-shoring and of the
rapidly growing dominance
of embedded systems development, also heavily using reconfigurable
platforms.
Instead of the de facto dominance (the almost-monopoly) of an
instruction-stream-based (von Neumann) mind set he advocates the early
introduction, already for freshmen, of a duality
of machine paradigms: instruction-stream-based and data-stream-based (von Neuann and anti machine), such that all graduates of
CS-related university programs should be qualified to cope with
hardware / configware / software partitioning decisions.
It is Hartenstein's achievement, that he receives an increasing number
of invitations to give talks on this complex of topic areas.(see erecent
talks).
- R.
Hartenstein (keynote): Reconfigurable HPC: torpedoed by Deficits in
Education ? ; Workshop
on
Reconfigurable Systems for HPC (RHPC); July 21,
2004, to be held in conjunction with HPC Asia 2004, 7th
International Conference on High Performance Computing and Grid in Asia
Pacific Region, July 20 - 22, 2004, Omiya Sonic City, Tokyo Area,
Japan (ppt) (pdf)
- R. Hartenstein (keynote): The Changing Role of
Computer Architecture Education within
CS Curricula; Workshop on
Computer Architecture
Education (WCAE 2004)
June 19, 2004, at 31st
International Symposium on Computer Architecture (ISCA), Munich, Germany, June 19-23, 2004 (ppt) (pdf)
*) CSG = Computer Structures Group
(Rechnerstrukturen)
at University of Kaiserslautern






©
Copyright 2001, University of Kaiserslautern, Kaiserslautern,
Germany
Webmaster