Prof. Mahmut Taylan Kandemir
The research at Penn State focuses on application modernization targeting emerging many-core systems. In particular, we investigate custom application mapping and optimization strategies that take into account unique architectural features of Intel® Xeon Phi™ coprocessors. Our research spans code and data optimizations (both application level and compiler level), latency hiding techniques such as prefetching, studying the tradeoffs between native and offload operation modes when running HPC applications, application restructuring techniques that target not only cache misses but also miss latencies, and investigation of novel architectural features that can be critical for future manycore systems (beyond Xeon Phi™ coprocessors). Our research will lead to (1) identification of necessary steps required for extracting maximum performance from Xeon Phi™ coprocessors during application mapping/optimization; (2) understanding whether conventional compiler optimizations as well as our new optimizations are effective when targeting large HPC codes, and if not, why; (3) identification of additional problems when mapping applications to a multi-Xeon Phi™ coprocessor system; and (4) comparison of our results against those that will collected using available commercial GPU systems. This research is also tightly coupled with a series of parallel programming and computing courses currently being taught at Penn State.