Obituary: Cleve Moler

Cleve Barry Moler, a pioneer in applied mathematics and computational science, passed away on May 20, 2026, at the age of 86. With Cleve’s passing, our field and all who had the privilege of knowing him share a deep loss. Cleve was indeed a gentle giant: brilliant, generous, and kind. He was a towering figure in numerical computing, a trusted colleague, and a dear friend. His contributions to the field of computing and the SIAM community will live on for many years to come. 

Through MATLAB, his best-known contribution to computing, Cleve made numerical methods accessible and interactive; his work at MathWorks empowered mathematicians, engineers, and students to build, test, and understand mathematical models with ease. What had once required difficult programming and deep knowledge of computing systems suddenly became available for quick and intuitive exploration. Cleve gave the computational science community a new language for discovery.

Cleve held professorships at the University of Michigan, Stanford University, and the University of New Mexico (UNM). He taught mathematics and computer science for nearly two decades; throughout those years, he remained committed to teaching students about the practical use of computation. His research and educational work grew from the conviction that numerical methods should be efficient, reliable, and available to all who need them.

That conviction led to multiple groundbreaking software contributions. Cleve was an author of EISPACK and LINPACK, two of the first libraries to make high-quality numerical linear algebra software available free of cost. EISPACK and LINPACK represented an important paradigm shift: rather than requiring individual researchers to write and debug their own numerical routines, the community could rely on portable, carefully designed, and well-tested software. The LINPACK Users’ Guide, which Cleve coauthored with Jack Dongarra, Jim Bunch, and Pete Stewart, set a standard of quality for documentation and dissemination of numerical software [1].

Cleve’s most widely recognized contribution began as an educational tool. While teaching at UNM, he wanted his students to be able to access the numerical power of LINPACK and EISPACK without first needing to become Fortran programmers. The result was the inaugural version of MATLAB (short for “matrix laboratory”). It provided students and researchers with an interactive environment where matrices, vectors, algorithms, and experiments could be explored easily, without the time and energy required to write and debug Fortran code. What began as a teaching aid became a new way to think about computation.

In 1984, Cleve co-founded MathWorks with Jack Little to continue developing and disseminating MATLAB. Since its inception—when PC-MATLAB was distributed by mail on floppy disks—MathWorks has grown into one of the most important institutions in technical computing. Cleve remained, until his passing, the intellectual and moral center of MATLAB’s evolution and delighted in interacting with and influencing the next generation of MATLAB developers.

MATLAB gave mathematical programmers easy, interactive access to graphics; complex arithmetic; matrices and tensors as objects; and built-in random numbers and constants (e.g., \(\pi\) and \(i\)). MATLAB quickly became part of the daily practice of countless people in industry and academia, reshaping how students learn linear algebra, how engineers develop designs, and how scientists test theories. Through MATLAB, Cleve’s approach to computing touches nearly every area where computation and mathematics meet. His judgment, taste, and deep understanding of numerical computation shaped the software at every level — not only its algorithms, but its spirit. 

In addition to his work in software, Cleve also contributed greatly to matrix algorithms. With Pete Stewart, he developed the QZ algorithm for the generalized eigenvalue problem, an important and enduring method for computing the generalized Schur decomposition of a matrix pencil. The QZ algorithm is now a fundamental tool in numerical linear algebra, with applications in stability analysis, control theory, and differential equations. It reflects Cleve’s characteristic style: deep mathematical insight combined with algorithmic practicality and the ability to foresee how methods would behave computationally. 

Beyond his mathematical skills, Cleve had a gift for exposition. His writing was clear, lively, and memorable, imbued with his talent for explaining complex numerical issues in an understandable way. He co-wrote Computer Methods for Mathematical Computations with George Forsythe and Michael Malcolm, a book that influenced generations of students and practitioners [2]. With Charlie Van Loan, he wrote the celebrated paper “Nineteen Dubious Ways to Compute the Exponential of a Matrix,” a classic work that combined humor, insight, and technical depth [3]. That paper went beyond reviewing methods for computing matrix exponentials to teach readers how to think critically about numerical algorithms, stability, conditioning, and the gap between mathematical formulas and reliable computation. 

Cleve was a pioneer in bringing numerical linear algebra into the era of distributed-memory parallel computing. While in the supercomputing division at Intel, he helped shape early thinking on the expression and implementation of dense matrix computations on massively parallel machines, where data had to be distributed across many processors and communication costs were as important as arithmetic speed. This work came at a formative moment, when scalable parallel computers were emerging but the necessary algorithms, software structures, and performance models to make them useful for scientific computing were still being invented. During this period, Cleve gave scientific computing one of its most memorable phrases: “embarrassingly parallel.” The term captured, with characteristic wit and clarity, a class of problems whose computations could be divided almost trivially among many processors, with little or no need for communication. Like so many of Cleve’s turns of phrase, it was both technically apt and instantly understandable.

Cleve’s work at Intel served as a bridge between classical numerical analysis and the emerging field of scalable high-performance computing. It reflected the same philosophy that animated much of his career: make powerful numerical methods usable on the machines of the day, in a way that opens the door for others. Whether through MATLAB, LINPACK, EISPACK, or his parallel computing work, Cleve helped turn advanced computation from a specialist’s craft into a practical tool for science and engineering.

Cleve’s accomplishments were recognized by some of the highest honors in computing, engineering, and applied mathematics. He was elected to the National Academy of Engineering in 1997 for the conception and development of widely used mathematical software. He received the Institute of Electrical and Electronics Engineers (IEEE) Computer Society Computer Pioneer Award in 2012 and the IEEE John von Neumann Medal in 2014. He was named a Fellow of the Computer History Museum in 2017, in recognition of the historic significance of MATLAB and his contributions to numerical computing. In 2023, he received the International Council for Industrial and Applied Mathematics Industry Prize for his extraordinary impact on mathematical and computational tools for science and engineering; he was then elected to the National Academy of Sciences in 2026.

Cleve’s relationship with SIAM was especially deep and lasting. He was a committed member of the SIAM community, served on the SIAM Board of Trustees, and served as President of SIAM from 2007 to 2008. He was named a SIAM Fellow in 2009 for his outstanding contributions to numerical analysis and software, including the invention of MATLAB. His involvement with SIAM reflected his belief that applied mathematics, numerical analysis, computer science, engineering, and industrial practice belong in close conversation. Cleve understood SIAM not merely as a professional society, but as a community — one built around ideas, service, collegiality, and the shared responsibility to make mathematics useful in the world.

Through his deep commitment to both SIAM and MathWorks, Cleve played a pivotal role in the creation of the MathWorks Math Modeling Challenge (M3 Challenge). Cleve served as a trusted bridge between SIAM and MathWorks; he championed the partnership and believed that MathWorks was the ideal collaborator for a program that aimed to inspire and empower the next generation of mathematical thinkers. For several years he attended the annual M3 Challenge Final Event in New York City as an honorary judge, where his enthusiasm for mathematics, technology, and education was evident. Students were delighted to meet the co-founder of MathWorks and the creator of MATLAB, which many of them used in their own work.

Cleve’s legacy is not only in software, algorithms, and institutions. It is also in the people he encouraged, challenged, and inspired. He was a colleague, mentor, and friend to many. He could ask a pointed question that went directly to the heart of the matter, but always with curiosity, humor, and goodwill. Those of us who saw him at conferences will remember that unmistakable booming voice, his careful attention to the technical details, and the warm smile that so often followed.

Cleve represented the best of our community: intellectual depth, practical insight, humility, and humanity. He helped build the foundations on which much of modern computational science rests, yet he remained approachable, generous, and deeply engaged with others. He had the rare ability to see both the mathematical essence of a problem and the practical needs of the person trying to solve it. That combination—rigor joined with usefulness, and brilliance joined with kindness—made him unique.

Cleve Moler leaves behind a remarkable legacy that will continue through the tools he created, the ideas he advanced, the company he helped build, and the many people whose work and lives he influenced. Cleve is survived by his wife, Patsy, his three daughters—Kathryn, Teresa, and Antonia —and his two stepdaughters, Carolyn and Byerly. We will miss his wisdom, his questions, his presence, and his insight. Most of all, we will miss our friend.

References
[1] Dongarra, J.J., Moler, C.B., Bunch, J.R., & Stewart, G.W. (1979). LINPACK users’ guide. Philadelphia, PA: Society for Industrial and Applied Mathematics.
[2] Forysthe, G.E., Malcolm, M.A., & Moler, C.B. (1977). Computer methods for mathematical computations. Upper Saddle River, NJ: Prentice Hall Publishing. 
[3] Moler, C.B., & Van Loan, C. (1978). Nineteen dubious ways to compute the exponential of a matrix. SIAM Rev., 20(4), 801-836.