I am a PhD student (admitted in 2007 and expected to graduate soon) in the Computer Science Department at the University of Maryland, College Park and work with Professor Ramani Duraiswami and Professor Nail A. Gumerov.

I received my B.S. degree in Mathematics from East China Normal University (China) in 2005 and my M.S. degree in Mathematics from National University of Singapore in 2007. At ECNU, I was working with Professor Xingzhi Zhan on matrix theory. My graduate supervisor at NUS was Professor Harald Niederreiter and my master thesis was on stream ciphers. 

Currently I am looking for full-time jobs in either research labs or industry. Here is my CV.

Research Interests:
My PhD dissertation is on scalable fast multipole methods (FMM) on heterogeneous architectures and GPU based scientific computing (such as fast kernel sums, matrix computations, iterative methods). Besides FMM related problems, I am also interested in stream ciphers, linear recurrence sequences, and matrix theory.

I am open to problems and opportunities to accelerate computations, applications or software performance by developing their parallel algorithms for GPUs, heterogeneous clusters or clouds.

Contact Information:
Email: huqi AT cs.umd.edu
A.V. Williams Building 3368, College Park, MD 20742


Publications (google scholar citation):
  • Refereed publications:
    • Q. Hu, N. A. Gumerov, and R. Duraiswami, "Scalable distributed fast multipole methods", 14th IEEE International Conference on High Performance and communications (HPCC-2012), Liverpool, UK, June 2012. (paper, slides, bibtex)
    • Q. Hu, N. A. Gumerov, and R. Duraiswami, "GPU accelerated fast multipole methods for dynamic N -body simulation", 24th International Conference on Parallel Computational Fluid Dynamics (ParCFD 2012) , Atlanta, USA, May 2012. (abstract; the full paper to appear in Computer & Fluids)
    • Q. Hu, N. A. Gumerov, and R. Duraiswami, "Scalable fast multipole methods on distributed heterogeneous architectures", in Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis, SC’11, (New York, NY, USA), pp. 36:1–36:12, ACM, November 2011. (Best Student Paper Finalist) (paper, slides, bibtex)
    • Q. Hu, M. Syal, N. A. Gumerov, R. Duraiswami, and J. G. Leishman, "Toward improved aeromechanics simulations using recent advancements in scientific computing", in Proceedings 67th Annual Forum of the American Helicopter Society, May 2011. (paper, slides, bibtex)
    • B. V. Srinivasan, Q. Hu, and R. Duraiswami, "GPUML: Graphical processors for speeding up kernel machines", Workshop on High Performance Analytics Algorithms, Implementations, and Applications, Siam Conference on Data Mining, April 2010. (paper, slides, bibtex)
    • Q. Hu, Y. Q. Li and X. Z. Zhan, "Possible number of ones in 0-1 matrices with a given rank", Linear and Multilinear Algebra, vol.53, no.6, pp. 435-443, 2005. (paper, bibtex)
  • Posters:
    • Q. Hu, N. A. Gumerov, R. Yokota, L. Barba and R. Duraiswami, "Scalable Vortex Methods Using Fast Multipole Methods", Technical Poster 2012 International Conference for High Performance Computing, Networking, Storage and Analysis, SC’12, Salt Lake City, UT, November 2012. (poster)
    • Q. Hu, N. A. Gumerov, and R. Duraiswami, "Heterogeneous Fast multipole methods Algorithm", February Fourier Talks 2012, College Park, MD, February 2012. (poster)
    • Q. Hu, N. A. Gumerov, and R. Duraiswami, "Fast N-body Algorithms for Dynamic Problems on the GPU", GPU Technology Conference, NVIDIA Research Summit, San Jose, CA, October 2010. (poster)
    • N. A. Gumerov, Y. Luo, R. Duraiswami, K. DeSpain, B. Dorland, and Q. Hu, "FLAGON: Fortran-9X Library for GPU Numerics", GPU Technology Conference, NVIDIA Research Summit, San Jose, CA, October 2009. (poster)
  • Others:
    • Q. Hu, "Fast Multipole Methods on Heterogeneous Architectures", Doctoral Showcase - Dissertation Research Showcase, 2012 International Conference for High Performance Computing, Networking, Storage and Analysis, SC’12, Salt Lake City, UT, November 2012. (acceptance ratio 25.5%)
    • Q. Hu, N. A. Gumerov, and R. Duraiswami, "Parallel algorithms for constructing data structures for fast multipole methods", under review for Journal of Parallel Computing. (arXiv)
    • B. V. Srinivasan, Q. Hu, N. A. Gumerov, R. Duraiswami, R.  Murtugutte, "Preconditioned Krylov slovers for kernel regressions", under review for Transactions on Pattern Analysis and Machine Intelligence.
Software:
  • Distributed heterogeneous FMM: this implementation can handle both Coulomb and Biot-Savart kernels, i.e., calculate potential, force, velocity and vortex streching terms. We are currently working on a new paper with fast streching term computations and the preliminary results can be found in our SC'12 poster.
  • GPUML (open source): it is a library that provides a C/C++ and MATLAB interface for speeding up the computation of the weighted kernel summation and kernel matrix construction on GPU.
  • FLAGON (open source): this is a middleware software to use GPUs via CUDA from Fortran 9X for Windows/Linux/MacOS.
Misc:
I am not only a super swimming fan but also play tennis, run and snowboard a lot.
I swim in Terrapin Masters Swim Club (also Santa Clara Swim Club last summer while I had my internship at NVIDIA) and spend most of my free time in the pool. Occasionally, I write blogs (in Chinese) to record some pieces of my life, but barely update it after I started to facebook and weibo. Sometime, I might post things related to CUDA, OS or computations into my tech blog.

The page access count: 
blog tracking




View Qi Hu's profile on LinkedIn 


                Last updated: 03 Nov 2012