Whether it be with Graphics Processing Units (GPUs) like those in our Blue Waters supercomputer, or Field-programmable Gate Arrays (FPGAs), Volodymyr Kindratenko's research has focused on pushing the limits of computational research, allowing high-performance computing architectures to be optimized to tackle unique and intensive challenges with great efficiency.
"Over the years, I have been able to contribute to the research of many of my colleagues, helping them to advance their state-of-the-art research by using experimental innovative specialized systems", stated Volodymyr Kindratenko. "From computational chemistry to chromodynamics in physics and even cosmology, we have worked to advance research using GPUs and novel algorithms."
Now, Volodymyr Kindratenko's work has taken him into one of the most rapidly-expanding realms of computer science: artificial intelligence (AI). Previously, artificial intelligence research has been limited by the capabilities of traditional computing systems, but new work from Volovymyr Kindtratenko and his team seeks to run AI applications at scale on high-performance computing systems, opening doors for potentials in machine learning at scales that were previously impossible.
"AI research is undergoing a major expansion due to the latest developments in complex, data-driven deep neural network training algorithms and high-performing computer systems tailored towards running these algorithms very efficiently. At NCSA, we are building a new computer system to run such algorithms at scale", stated Volodymyr Kindratenko. "This system is funded by the NSF MRI programme and is developed in collaboration with IBM and NVIDIA. Our objective is to bring to the UIUC campus a unique resource that will enable domain scientists to focus on their science problems rather than spend time optimizing their code and waiting for lengthy computations. This unique resource will significantly accelerate the AI research on campus."