The primary goals of this collaborative CRI grant include: (1) build Marcher, a power-measurable high-performance computing infrastructure equipped with Intel Many Integrated Cores, Nvidia GPUs, high-capacity memory systems and hybrid storage systems; (2) build easy-to-use power measurement tools for the Marcher system; (3) support numerous green computing research projects using the Marcher system; and (4) provide infrastructure support for advancing green computing education. Two universities (Texas State University and Clemson University) are involved in this collaborative grant (#1305359and #1551262).The proposed goals have been accomplished and the outcomes derived from this collaborative grant are summarized below:

1)  Research Activities: Numerous research projects were supported, including accurate power measurement for GPUs, energy consumption analysis of parallel databases, characterizing energy consumption of parallel programs running on Intel Xeon Phi, energy efficiency analysis of deep convolutional neural networks, energy-efficient scheduling for heterogenous high performance computing systems, energy analysis and optimization for resilient linear system, intelligent power coordination for power-bounded systems, memory optimization for green cloud, new metrics to evaluate software energy efficiency, exploring the impact of language, runtime, compiler, object-oriented programming and design patterns on software energy efficiency. These projects have generated several novel algorithms and new studies, which contribute to the green computing discipline.  

2)  Publications: By the time of submitting this report, 17 peer-reviewed papers have been published in highly recognized IEEE/ACM sponsored conferences/workshops, which include the IEEE Transactions on Computers, the Journal of Big Data Research, the IEEE International Green and Sustainable Computing Conference, the IEEE International Conference on Cluster Computing, the International Conference on Parallel Processing, the ACM International Conference on Architectural Support for Programming Languages and Operating Systems, the ACM International Conference on Big Data Science and Computing, the IEEE International Conference on Sustainable Computing and Communications, the IEEE International Conference on Networking, Architecture, and Storage. In addition, two papers (one journal and one conference) have been submitted and currently under review.

3)  Training: 7 graduate students and 22 undergraduate students participated in the aforementioned research projects led by the PIs (Zong, Ge, and Gu). These research projects helped both graduate and undergraduate students gain research experiences and interests in green computing. A couple of students made impressive achievements. Sarah Abdulsalam, a female graduate student at Texas State University, published two conference paper as the first author and was hired by Intel. Divya Mahajan, a female graduate student at Texas State University, published a conference paper as the first author and was hired by Dell. Stuart Olsen, an undergraduate student at Texas State University, published a conference paper as the first author and was hired by Google. Sang Mercado, an undergraduate student at Texas State University, was hired by LinkedIn.     

4)  Education: The Marcher system and research findings derived from this collaborative grant have been integrated into various levels of classes taught by PIs at two institutions. Specifically, PI Zong created three green computing courses (CS7333 for Ph.D. students, CS5369Y for master students, and CS4379Y for undergraduate students) at Texas State University. Over 80 students from these courses have utilized the Marcher system to gain first-hand experience and intuition in the fundamental concepts and research frontiers of green computing. PI Ge created a graduate course (CPSC8810) at Clemson University where over 30 Masters and Ph.D. students worked on related advanced course projects. 

5) Broad Impact: The PIs strive to attract minority and female students involved in the research projects supported by this grant. 75% of supported graduate students at Texas State University are female. Sarah Abdulsalam, a female graduate student at Texas State University, is one of the exemplary minority students. She worked with PI Zong for less than two years and published two conference papers as the first author. She was hired by Intel after her graduation. Moreover, there are two female undergraduate students and two Hispanic undergraduate students were supported by this grant. Two REU students who worked with PI Ge have gone to or are applying for graduate school. In addition, the Marcher system has been used by researchers/faculty from the Texas Advanced Computing Center, the Pacific Northwest National Lab, Washington State University, North Carolina State University, University of Texas at San Antonio, and University of North Texas to support various research and education projects. Last but not the least, the Greensoft website was created to provide detailed documentation and case studies on how to use the Marcher system, which can help grow the community conducting research and education in green computing.

6)  Released Data: The research findings of this project are published in conferences or workshops for public knowledge. The Greencode platform serves as a free platform to support research and education in green computing. To date, it has processed approximately 20,000 code submissions from hundreds of users. Users can share their energy efficient programs with the community for free.   

Last Modified: 09/04/2018
Modified by: Ziliang Zong