Oyekunle Ayinde "Kunle" Olukotun is a British-born Yoruba Nigerian computer scientist. He holds the title of Cadence Design Systems Professor at the Stanford School of Engineering and is a professor in the departments of Electrical Engineering and Computer Science at Stanford University. He also leads the Stanford Pervasive Parallelism Lab. Olukotun is known as the "father of the multi-core processor" and directed the Stanford Hydra Chip Multiprocessor research project. His achievements include designing the first general-purpose multi-core CPU, developing single-chip multiprocessor and multi-threaded processor designs, and helping create multicore CPUs and GPUs, transactional memory technology, and domain-specific languages programming models. His research focuses on computer architecture, parallel programming environments, scalable parallel systems, domain-specific languages, and high-level compilers.

Education

Olukotun completed his bachelor's degree at Calvin College in Grand Rapids, Michigan. He received a master's degree in 1987 and a doctorate in 1991 from the University of Michigan, both in Computer Science and Engineering. His advisor during these studies was Trevor Mudge.

Career

In 1991, Olukotun joined Stanford's Department of Electrical Engineering. At Stanford, he led the Stanford Hydra chip multiprocessor (CMP) research project, which helped create multiprocessors that supported thread-level speculation. In 2000, he started Afara Websystems, a company that designed and built high-throughput, low-power processors for server systems using chip multiprocessor technology. In 2002, Afara Websystems was bought by Sun Microsystems. The Afara multicore processor called Niagara, created by Olukotun, was acquired by Sun. Niagara-based processors are now used in all Oracle SPARC-based servers and have generated billions of dollars in revenue. During his time at Sun, Olukotun was one of the main designers of the 2005 UltraSPARC T1 processor.

In 2017, Olukotun and Chris Ré started SambaNova Systems. SambaNova Systems is working on a new computing platform to support machine learning and data analytics. Currently, Olukotun leads the Stanford Pervasive Parallelism Lab, which aims to make heterogeneous parallel computing easier to use. He is also part of the Data Analytics for What’s Next (DAWN) Lab, which is creating tools to support practical machine learning.

Research

Olukotun's research areas include computer architecture, ways to write programs that use multiple processors at once, systems that can handle many tasks simultaneously, specialized programming languages, and tools that convert high-level code into machine language.

He leads the Stanford Hydra chip multiprocessor (CMP) research project, which changed the way computers work by making multi-core processors available to regular users and high-end systems.

In the mid-1990s, Olukotun and his co-authors said that multi-core processors would use hardware more efficiently than other types of processor designs.

In 2008, Olukotun returned to Stanford and started the Pervasive Parallelism Laboratory after receiving $6 million in funding from several computer-industry companies. His recent work focuses on specialized programming languages that allow algorithms to work on different kinds of parallel hardware, such as multi-core systems, graphics processing units, and field-programmable gate arrays.

Olukotun serves on the board of advisors for UDC, a Nigerian venture capital firm. He was chosen as a Fellow of the Association for Computing Machinery in 2006 for his "contributions to multiprocessors on a chip and multi-threaded processor design." He became a Fellow of the IEEE in 2008.

Olukotun uses words from his Yoruba heritage in his work. The name of the company he founded, Afara, means "bridge" in Yoruba. He also named his server at Stanford Ogun, which is the name of a Yoruba god of iron and steel. This is a play on words because large computers are often called "big iron."

Olukotun directs the Stanford Pervasive Parallelism Lab (PPL), which encourages the use of parallelism in many application areas. He is also a member of the Data Analytics for What's Next (DAWN) Lab.

Olukotun holds 12 U.S. patents. He has published more than 150 scientific papers and wrote two textbooks.

Awards and honors

• Received the Eckert–Mauchly Award in 2023
• Was named a member of the National Academy of Engineering in 2021
• Received the IEEE Computer Society Harry H. Goode Award in 2018
• Received the Michigan Engineering Alumni Merit Award in 2017
• Was named an ACM Fellow in 2006

Books

• S. W. Keckler, K. Olukotun, and H. P. Hofstee, Multicore Processors and Systems, published by Springer Publishing Company, Inc. in 2009.
• K. Olukotun, L. Hammond, and J. Laudon, Chip Multiprocessor Architecture: Techniques to Improve Throughput and Latency, published as a Synthesis Lecture on Computer Architecture by Morgan Claypool Publishers in 2007.