Sayantan Sur Department of Computer Science Email: surs@cse.ohio-state.edu 395 Dreese Laboratories, Web: http://www.cse.ohio-state.edu/~surs/ 2015 Neil Avenue, Phone: 614-804-9898 (Mobile), 614-292-8501 (Off) Columbus OH 43210 Fax: 614-292-2911 Objective To obtain a challenging full-time position in the field of Network-Based Computing in Research Labs or Industry. Research Interests High-speed interconnects, InfiniBand Architecture, User-level communication protocols, Storage systems, Parallel I/O systems, Parallel Architectures Education PhD. Student Current GPA: 3.6 Department of Computer Science The Ohio State University, Columbus, Ohio 2002 - July 2007 (est.) Bachelor of Technology (with Honors) 76% (Top 5% of class) Electrical and Electronics Engineering National Institute of Technology, Calicut, India July 2001 Work Experience Graduate Research Assistant Network-Based Computing Laboratory Department of Computer Science 2003 - present * Lead role in overall architecture of MVAPICH (MPI over InfiniBand) Summer Intern Argonne National Laboratory Mathematical and Computer Sciences Division June-September 2006 * Design and implementation of InfiniBand network channel for MPICH2/Nemesis communica- tion layer Summer Intern Intel Corp. Parallel and Distributed Software Division June-September 2005 * Advanced Semantics for message-based interconnects and transports Graduate Teaching Assistant Department of Computer Science The Ohio State University 2002 - 2003 * Lab instructor for CSE 200 and grader for CSE 655 (Principles of Programming Languages) Member of Technical Staff Sun Microsystems Bangalore, India 2001 - 2002 * Design and implementation of automation framework for iPlanet WebMail and Address Book Awards * Outstanding Research Award - Department of Computer Science and Engineering * Student Travel Award - ACM Symposium on Principles and Practice of Parallel Programming (PPOPP - 2006) Publications Journal Articles * W. Yu, S. Sur, D. K. Panda, R. T. Aulwes and R. L. Graham. "High Performance Broadcast Support in LA-MPI over Quadrics", International Journal of High Performance Computer Applications 2005. Conference Publications * S. Sur, M. Koop and D. K. Panda "High-Performance and Scalable MPI over InfiniBand with Reduced Memory Usage: An In-Depth Performance Analysis", SuperComputing 2006. * S. Sur, L. Chai, H.-W. Jin and D. K. Panda "Shared Receive Queue Based Scalable MPI Design for InfiniBand Clusters", International Parallel and Distributed Processing Symposium 2006. * S. Sur, H.-W. Jin, L. Chai and D. K. Panda "RDMA Read Based Rendezvous Protocol for MPI over InfiniBand: Design Alternatives and Benefits", Symposium on Principles and Practice of Parallel Programming 2006. * S. Sur, U. Bondhugula, A. Mamidala, H.-W. Jin and D. K. Panda "High Performance RDMA Based All-to-all Broadcast for InfiniBand Clusters", International Conference on High Performance Computing 2005. * S. Sur, A. Vishnu, H.-W. Jin, W. Huang and D. K. Panda "Can Memory-Less Network Adapters Benefit Next-Generation InfiniBand Systems? ", Hot Interconnects Symposium 2005. * H.-W. Jin, S. Sur, L. Chai and D. K. Panda "LiMIC: Support for High-Performance MPI Intra-Node Communication on Linux Cluster ", International Conference on Parallel Processing 2005. * L. Chai, S. Sur, H.-W. Jin and D.K. Panda, "Analysis of Design Considerations for Optimizing Multi- Channel MPI over InfiniBand ", Workshop on Communication Architecture for Clusters (CAC), to be held in conjunction with IPDPS 2005. * S. Sur, H.-W. Jin and D.K. Panda, "Efficient and Scalable All-to-All Exchange for InfiniBand-based Clusters", International Conference on Parallel Processing 2004. * W. Yu, S. Sur, D. K. Panda, R. T. Aulwes and R. L. Graham. "High Performance Broadcast Support in LA-MPI over Quadrics", Los Alamos Computer Science Institute Symposium 2003. Technical Reports * K. Vaidyanathan, S. Sur, S. Narravula, P. Sinha. "Data Aggregation Techniques in Sensor Net- works", OSU-CISRC-11/04-TR60. Presentations * Shared Receive Queue based Scalable MPI Design for InfiniBand Clusters: Int'l Parallel and Dis- tributed Processing Symposium (IPDPS), 2006. * Efficient SMP-Aware MPI-Level Broadcast over InfiniBand's Hardware Multicast: Communication Architecture for Clusters (CAC) Workshop, Held in conjunction with Int'l Parallel and Distributed Processing Symposium (IPDPS), 2006. * RDMA Read Based Rendezvous Protocol for MPI over InfiniBand: Design Alternatives and Benefits: Symposium on Principles and Practice of Parallel Programming (PPOPP), 2006. * InfiniBand - Designing HPC Clusters, Storage/File Systems, and Datacenters: Tutorial Co-presenter at Super Computing 2005. * Efficient and Scalable All-to-all Exchange for Infiniband based Clusters: International Conference on Parallel Processing 2004. Selected Projects MVAPICH (MPI over InfiniBand) * Lead role in evolving the overall architecture as the InfiniBand technologies mature. Optimized point-to-point communication * Designed an optimized rendezvous protocol which uses RDMA-Read and asynchronous progress mechanisms to provide better communication and computation overlap. * Designed a buffering scheme which exploits Shared Receive Queues for improving memory scalability on multi-thousand InfiniBand clusters. * Investigating protocols for sharing RDMA buffers across various connections to achieve higher buffer utilization. Design and implementation of high performance collectives using RDMA * Designed and implemented a hybrid mechanism for an efficient RDMA based Alltoall and Allgather over InfiniBand. Optimized shared memory message passing implementation * Kernel based mechanism for implementing very efficient message passing by reducing several copies. * Processor local shared memory buffers and message queues for NUMA architectures. Open MPI * Design and implementation of point-to-point transport layer (PTL) over InfiniBand VAPI. * Designed an Out-of-Band (OOB) mechanism to set up dynamic InfiniBand connections. NIC-based Atomic Operations * Modified the GM low level message passing firmware to allow host based memory operations be atomic with respect to remote NIC-based atomic operations. The Peterson's algorithm was used to maintain coherency. Data Aggregation in Sensor Networks * Designed new data aggregation schemes in sensor networks to allow efficient utilization of energy. Simulation results were obtained over NS-2. Micro Secure * Designed and developed a PC interfaced Microcontroller (Intel 80C196) based security and power control system. Skills C, C++, Java, User-Level network (InfiniBand, GM) programming, Kernel level programming on Linux, Design of multi-threaded software, Shell scripting, Latex. Professional Activities * Reviewer for papers submitted to: International Parallel and Distributed Processing Symposium (IPDPS) 2004, 2005 International Symposium on Performance Analysis of Systems and Software (ISPASS) 2005 International Conference on Communications (ICC) 2005 International Journal of Computers and Applications * Member of Upsilon Pi Epsilon Computer Science Honor Society. * ACM student member. References * Dr. D. K. Panda, The Ohio State University email: panda@cse.ohio-state.edu