BIO
CV (PDF)
RESEARCH

 

Naga K. Govindaraju
Senior Researcher
Many-core Technology Incubation
Microsoft Corporation
E-mail: nagag at microsoft dot com
Phone: (425)-421-6919



SOFTWARE:

A high performance DFT library on GPUs using the modern compute languages on GPUs with performance up to 300 GFLOPS on a single GPU. Library handles non-power-of-two sizes, multi-dimensional transforms and achieves high numerical accuracy on single-precision inputs. Joint work with Brandon Lloyd, Yuri Dotsenko, Burton Smith and John Manferdelli at Microsoft. Research will be presented at ACM/IEEE SuperComputing 2008.

GPUFFTW: High-Performance Power-of-Two FFT Library using Graphics Processors is released publicly on May 26, 2006. The library is highly tuned for GPUs using compute and memory optimizations and is able to achieve over 15 GFLOPS on large 1D FFTs using a NVIDIA 8800 GPU. More than 4000 download requests at the release and the technical paper appeared in ACM SuperComputing 06 (joint work with Scott Larsen, Jim Gray and Dinesh Manocha).

GPUTeraSort (joint work with Jim Gray, Ritesh Kumar and Dinesh Manocha) is a high performance-price solution for handling hundred gigabyte scale databases! CPU-based sort speeds have plateaued during the past decade (see Jim's commentary). GPUTeraSort shows a significant performance benefit achieving 3x more records/sec. GPUTeraSort is the world's fastest reported performance/price sorting system and is the 2006 Indy PennySort winner. Press releases in NYTimes and Wired Magazine

LUGPULIB : A Fast Dense Linear Solver Library on GPUs released publicly in November 2005. Over 1000 download requests were received.

GPUSORT : A High Performance GPU Sorting Library released publicly in July 2005. Over 3500 download requests were received.


RECENT COURSES/TUTORIALS IN MAJOR CONFERENCES:

GPGPU: General Purpose Computation on Graphics Hardware,
co-organizer with M. Houston,
ACM SIGGRAPH 2007 Tutorials

Query Co-Processing on Commodity Hardware,
co-organizer with A Ailamaki, S Harizopolous, D Manocha,
VLDB 2006 Tutorials

Query Co-Processing on Commodity Hardware,
co-organizer with A Ailamaki, D Manocha,
IEEE International Conference on Data Engineering 2006 Tutorials

GPGPU: General Purpose Computation on Graphics Hardware,
with I Buck, M Harris, M Houston, A Lefohn, D Luebke, J Owens, M Segal,
ACM SuperComputing 2006 Tutorials

Collision Handling in Dynamic Simulation Environments,
with M Teschner, B Hiedelberger, D Manocha, G Zachmann, S Kimmerle, A Fuhrmann,
Eurographics 2005 Tutorials

GPGPU: General Purpose Computation on Graphics Hardware,
with I Buck, M Harris, J Kruger, A Lefohn, D Luebke, T Purcell, C Wooley,
ACM SIGGRAPH 2005 Courses

PROFESSIONAL SERVICES:

Program Committee Member for Eurographics 2009
Program Committee Member for ACM I3D 2009
Program Committee Member for Very Large Databases (VLDB) 2009
Program co-chair and Program Committee Member, ACM Graphics Hardware 2008
Program Committee Member for Parallel Architectures and Compilation Techniques (PACT) 2008
Program Committee Member for Very Large Databases (VLDB) 2008
Program Committee Member for ACM I3D 2008
Session Chair, ACM I3D 2007
Industrial Liason Co-chair for ACM Virtual Reality and Software Technology, 2007
Co-organizer, ACM SuperComputing Multi- and Many-core Workshop 2007
Program Committee Member for First International Workshop on General Purpose Computing on Graphics Processing Units 2007
Program Committee Member for Pacific Graphics 2007
Program Committee Member for Graphics Hardware 2007
Program Committee Member for ACM I3D 2007
Program Committee Member for ACM Virtual Reality and Software Technology, 2006
Program Committee Member for ACM Workshop on Data Management on New Hardware 2006
Program Committee Member for Graphics Hardware 2006
Program Committee Member for Pacific Graphics 2005