by David Goddard
The Innovative Computing Laboratory was founded at UT in 1989 as a way to foster collaboration, push boundaries, and develop new methods, libraries, and systems, all in relation to the field of supercomputing.
For more than 30 years, Distinguished Professor of Computer Science Jack Dongarra guided the ICL as it took on challenge after challenge, helping it become a driving force behind the understanding of high-performance computing that exists today.
Now, with Dongarra having transitioned to professor emeritus status, recently appointed Director Hartwig Anzt has a new idea for the ICL and what he hopes it can become.
“There lies enormous potential in bringing our HPC expertise into machine learning and artificial intelligence research,” said Anzt, who also took on the role as an associate professor of EECS as part of the change. “I hope to succeed in teaming up with ML/AI experts to create a joint effort while also increasing the interaction with the HPC and computer science groups at ORNL.”
Anzt is familiar with the ICL, having spent part of his doctoral research working there after meeting Dongarra at a 2010 conference in Iceland and running the idea by him.
He late continued his relationship with the lab, first as a postdoc and later as a research scientist.
After an early career award from the Helmholtz Association gave him the ability to start his own research group at the Karlsruhe Institute of Technology in Germany, he became a research consultant with ICL, a position he continued to hold until he became director.
“I wanted to give something back,” he said. “Over the years, I have benefited a lot from ICL. I would not have been able to build my career, build my group in Germany, without all what I learned from ICL and the vision I created while working here.”
Anzt said it came as shock to both people who worked with the ICL and to the larger computing community when Dongarra announced his intention to retire.
Despite the initial surprise and wonderment at how the ICL could survive without Dongarra, Anzt said, he had a vision for what it could become. And his goal is now to realize that vision.
Helping him achieve that goal is the lab’s reputation with partners and facilities around the world, including several national labs and well-known titans of the computing industry. That shared history will be beneficial, especially in the beginning of Anzt’s tenure, as the trust that ICL has built will allow a smooth process moving forward.
“In collaborative software development, one needs to trust another to share ideas and code, and as collaborative software development is a major activity in ICL, this trust is critical,” said Anzt. “I hope to continue the nationwide and international network of collaboration and research exchange.”
Anzt said another important factor in ICL’s success is that it is blessed with an outstanding administration and technical support group that understands that ICL can only succeed if the researchers can focus on research.
Citing his own journey, he also stressed the importance of supporting graduate students and young researchers, including financial support across borders.
“I understand that this requires financial resources and is a financial investment with unclear outcome, as not every student will come back and stay in academia,” he said. “Nevertheless, I think it is the best investment we can make.”
More than just investing in students, that support is an investment in the continued success of the ICL—for now, and for the next 30-plus years.