Mahmut Taylan Kandemir
Kandemir’s research spans optimizing compilers, runtime systems, embedded systems, I/O and high performance storage, and power-aware computing. His research has been/is funded by the National Science Foundation, DARPA, ONR, and Microsoft Corporation. He has completed 13 funded research projects (including one supported by his NSF CAREER AWARD) and is currently PI or co-PI on 11 funded projects. His contributions to optimizing compilers include several novel analysis and optimization techniques targeted at eliminating power and memory bottlenecks during program execution. His recent compiler research focuses on emerging multi-core architectures and is oriented toward providing multi-core specific programming support and optimizations, which are seen by both industry and academia as the most critical problem that could prevent the wide-adoption of future multi-cores. His research in multi-cores also includes runtime support that helps the architecture adapt to dynamic variations in resource availability and application characteristics. Kandemir’s contributions to I/O include file and runtime system optimizations for disk accesses, I/O prefetching and storage caching. In collaboration with Argonne National Lab, he is transitioning the research ideas developed by his group to product file systems, whose performance is critical for many large-scale data intensive simulation codes of national importance, ranging from astrophysics to bio-informatics to computational chemistry. In the domain of embedded computing, Kandemir and his students have contributed to the design of custom high-level memory hierarchies and placed a tool recently into public domain. His research efforts on power-aware computing have focused on design of power estimation models and tools and resulted in holistic solutions that include both hardware and software. Kandemir continuously collaborates with researchers from other engineering and science disciplines, international scholars as well as people from industry such as Microsoft, HP Labs, Intel and NVIDIA. Kandemir’s research has so far resulted in over 80 journal publications and over 300 conference/workshop papers. His papers appear in premier programming language and compiler conferences (POPL, PLDI), design automation/embedded system conferences (DAC, ICCAD, LCTES), system conferences (ASPLOS, SIGMETRICS), high performance computing and I/O conferences (FAST, Supercomputing, ICS, IPDPS) and architecture conferences (ISCA, MICRO, HPCA), as well as in ACM and IEEE transactions. Kandemir is a recipient of the Outstanding Research Award of the Penn State Engineering Society (PSES). In the last two years, two of his publications have received the best paper awards (ICPADS in 2006 and IPDPS in 2008); and three of his publications have been nominated for the best paper awards in the last four years. One of his papers has been the top download from ACM’s Digital Library in October 2006.
Kandemir’s teaching has focused on integrating research into both undergraduate and graduate curriculum. He regularly teaches undergraduate programming language and compiler courses and the graduate level compiler course. Kandemir has also taught several new courses at the graduate level at Penn State, including High Level System Design, Optimizing Compilers, and Programming Chip Multiprocessors. Kandemir has graduated 11 Ph.D. students and 8 master students so far; four of his Ph.D. graduates are now faculty members at other universities (both in the US and outside). He is currently supervising 15 Ph.D. students, four of whom are co-advised, and one master student.
Kandemir is currently serving as the Graduate Program Coordinator for the computer science and engineering department at Penn State. This program currently has more than 250 graduate students and is one of the largest programs in the college of engineering. He has served in several departmental committees such as Curriculum Committee, Personnel Committee, and Graduate Committee. He also served in the Engineering Advising Center at Penn State and on the Graduate Council. Kandemir has served / is serving in the capacity of program chair, co-chair and program committee member of several conferences and workshops in his research fields. He currently serves as an associate editor for two journals (ACM TODAES since 2004 and IJDSN since 2006). He has received a certificate of appreciation for outstanding professional service from IEEE.