Kishore Kumar Pusukuri

Kishore's Photo

Kishore Kumar Pusukuri

PhD Candidate
Department of Computer Science and Engineering
University of California, Riverside
e-mail: kishore@cs.ucr.edu

Research interests : performance issues in multicores, operating systems, run-time techniques, cloud computng, machine learning, and power-aware computing.

Prof. Rajiv Gupta is my advisor.


Research Experience

  • UC Riverside, Riverside, CA. Research Assistant. [Sept. 2007 - Sept. 2012]
    Run-time support for exploiting multicore systems. (Dissertation Topic)
    Since multicore systems offer greater performance via parallelism, future computing is progressing towards use of machines with large number of cores. However, due to the complex interaction among characteristics of multithreaded applications, operating system policies, and architectural characteristics of multicore systems, delivering high performance on multicore systems is a challenging task. This dissertation addresses the above challenge by developing runtime techniques to achieve high performance when running a single multithreaded application as well as high system utilization and fairness when running multiple multithreaded applications. The runtime techniques are based on a simple monitoring system that captures important application characteristics and relevant architectural factors with negligible overhead.
  • Sun Microsystems Labs, Menlo Park, CA. Research Intern. [June 2008 - Sept. 2008]
    Developed a methodology for developing simple and robust power models using performance monitoring events for multicore systems running OpenSolaris . The basic idea is correlating power consumption of a benchmark program with its performance. By using applicable model selection and model assessment techniques, we developed a simple and robust power model, which was shown to predict the power consumption with better than 95% prediction accuracy.
  • Sun Microsystems Labs, Menlo Park, CA. Research Intern. [June 2009 - Aug. 2009]
    Developed FACT, a Framework for Adaptive Contention-aware Thread migrations, which measures the relevant performance monitoring events online, learns to predict the effects of interference on performance of workloads using supervised learning techniques, and then makes optimal thread scheduling decisions.
  • Intel, Hillsboro, OR. Graduate Technical Intern. [June 2011 - Sept. 2011]
    Analysed performance of database transaction applications running on machines with a large number of cores running Linux, and proposed OS level optimization and scheduling techniques to improve their performance.

Publications

  • Kishore Kumar Pusukuri, Has One-thread-per-core Binding Model Become Obsolete for Multithreaded Programs on Multicore Multiprocessor Systems?. In proceedings of USENIX HotPar’13, San Jose, USA, June 2013.

  • Kishore Kumar Pusukuri, Rajiv Gupta, Laxmi N. Bhuyan, ADAPT: A Framework for Co-scheduling Multithreaded Programs. In ACM Transactions on Architecture and Code Optimization (ACM TACO), 2013.

  • Kishore Kumar Pusukuri, Rajiv Gupta, Laxmi N. Bhuyan, Thread Tranquilizer: Dynamically Reducing Performance Variation. In ACM Transactions on Architecture and Code Optimization (ACM TACO), Volume 8 Issue 4, 2012.

  • Kishore Kumar Pusukuri, Rajiv Gupta, Laxmi N. Bhuyan, Thread Reinforcer: Dynamically Determining Number of Threads via OS Level Monitoring. In proceedings of IEEE International Symposium on Workload Characterization (IISWC), Austin, Texas, USA, Nov. 2011. (pdf)

  • Kishore Kumar Pusukuri, Rajiv Gupta, Laxmi N. Bhuyan, No More Backstabbing... A Faithful Scheduling Policy for Multithreaded Programs. In proceedings of the Twentieth International Conference on Parallel Architectures and Compilation Techniques (PACT), Galveston Island, Texas, USA, Oct. 2011. (pdf)

  • Kishore Kumar Pusukuri, David Vengerov, Alexandra Fedorova, Vana Kalogeraki, FACT: a Framework for Adaptive Contention-Aware Thread Migrations. In proceedings of ACM International Conference on Computing Frontiers (CF), Ischia, Italy, May 2011. (pdf)

  • Kishore Kumar Pusukuri, David Vengerov, Alexandra Fedorova, A Methodology for Developing Simple and Robust Power Models using Performance Monitoring Events. In proceedings of WISOCA, Austin, Texas, USA, June 2009. (pdf)

  • Gaurav Dhiman, Kishore Kumar Pusukuri, Tajana Rosing, Analysis of Dynamic Voltage Scaling for System Level Power Management. In proceedings of HotPower, San Diego, CA, USA, Sept. 2008. (html)

  • Kishore Kumar Pusukuri, Atul Negi, Applying machine learning techniques to improve GNU/Linux process scheduling, In proceedings of IEEE International Tencon Conferece'05, Melbourne, Australia, Dec. 2005. (pdf)

  • Kishore Kumar Pusukuri, Atul Negi, Characterizing process execution behaviour using machine learning techniques. International Workshop on Cluster/Data Center Dynamic Provisioning and Resource Management, IEEE HiPC, Bangalore, India, December 2004. (pdf)

Others

  • Alva L. Couch, Kishore Kumar Pusukuri. Report on the workshop on Power Aware computing and Systems (HotPower'08). login, The Usenix Magazine, volume 34, Number 2, pages: 110 - 114, April 2009.