IN MEMORY OF

Peter Kornerup

Peter Kornerup Peter Kornerup

Peter Kornerup was born in Aarhus, Denmark in 1939. He matriculated at Aarhus with studies in mathematics but soon moved to the new field of computer science. In 1971 he became chairman of the computer science department at Aarhus University a position he held through 1987. In 1988 he was appointed a full professor at Odense University, where he has been emeritus since July 2009. He has spent periods of research at the University of Southern Louisiana, Stanford University, Southern Methodist University in Dallas, and ENS Lyon.

Peter's research was focused throughout his 40+ year academic career on computer arithmetic in the broadest sense. This encompassed hardware design, programming language interpretation of arithmetic representations and operations, and the mathematical foundations of number systems and computer arithmetic. From 1971-73 with Bruce Shiver in Aarhus he built a companion computer to an existing mainframe that would process arithmetic operations in a variety of formats, thus implementing a pioneering "mathematical co-processor." The arithmetic "operation type" language extensions and hardware architecture of the system were presented in several of his papers in our early 1975 3rd ARITH proceedings, which also introduced our familiar logo.

This same year Bruce introduced me to Peter having noted our parallel interests in number representations and mathematical formulations of arithmetic operations including precise rounding. What followed was a cherished 35 year collaboration I had with Peter refining finite precision rational, residue, and radix number systems and their arithmetic. His contributions to our joint papers were seminal, introducing measured redundancy to achieve faster algorithms in hardware. Many of our papers regularly appeared first in ARITH proceedings, with our joint research then consummated in our 2010 book "Finite Precision Number Systems and Arithmetic". In Europe Peter furthered his research and student mentoring with a long collaboration in Lyon with Jean-Michel Muller and his institute, producing many affiliations of lasting value to our ARITH community.

Professor Kornerup's career in academia achieved more than research accolades. He was an efficient and fair minded administrator. Repeatedly he was voted to be chairman and sometimes dean. Over eleven years, he built Aarhus's computer science department into one of the strongest departments in Europe, a position still maintained based on his focused foundation of mathematical rigor and breadth of coverage (including computer architecture). His contributions to academia in the broad sense included his advocacy of computer architecture as integral to computer science in its foundation as a discipline for the established academies, not just technical institutes. He participated in inter-European visits and committees to broaden computer science curricula and promote its impending universal importance throughout human society. His fairness and thoughtful evaluation of student contributions made him a regular evaluator of examinations throughout the Danish university system and beyond. This was the same intensive evaluation and fair minded incisive critiques he was known for by us who have seen and heard his evaluations of ARITH and IEEE Transactions submissions.

Personally Peter was a man of strong character and morality. His honesty, humility, collegiality and faithfulness to family and traditions have remained steadfast as his reputation and influence have grown over the years. At conferences and in person he was generous and took pride only in introducing Danish traditions to others that visited him.

His departure to a full professorship in Odense was personal tribute to his wife Margot's new found position as a judge in Odense. His ties to country life included maintaining his summer house outside Aarhus where he could nurture his workshop and outdoor grilling passions. He was a wonderful host, cooking and serving with his wife, often using spices he picked in the wild.

Indeed Peter was a role model in his profession and as both colleague and friend. He will be sorely missed.

David W. Matula

 


Learning that Peter passed away was very sad news. When I was a young researcher, two more senior persons helped me much, in a quasi-paternal way: one of them was Peter. Our first professional cooperation goes back to 1991, when Peter was Program chair of the Grenoble Arith symposium and I was local organiser. And our last cooperation is very recent: a paper published in march 2017 in the journal Mathematics of Computation, probably Peter's last article. But more important that professional work is friendship. In this world full of people with oversized ego, it was so refreshing to share moments with this quiet, modest, friendly, honest person. Peter, we all miss you.

Jean-Michel Muller

 


Peter Kornerup will be remembered as a significant contributor to the field of computer arithmetic. A mathematician who also worked on computer science problems, he brought a unique mix in his research in computer arithmetic. His contributions to number representations, arithmetic algorithms, and designs are characterized by a pervasive mathematical rigor. He participated in organizing our symposia since early days as chair and as a member of possibly all program committees, including the steering committee. He was a meticulous reviewer contributing to the quality of our conferences. It was rewarding to listen to his comments and technical arguments - always demonstrating a deep understanding and broad knowledge. Talking with him was always pleasant, not only about arithmetic. He was curious and interesting. I am grateful that I have known Peter for more than 40 years - a memory that will always stay with me.

Milos Ercegovac