Monday
October 26
4:00 - 4:50 PM 
Kelley 1001 [map]
<http://oregonstate.edu/cw_tools/campusmap/?&offsetX=1576&offsetY=526&po
int=?298,111> 

 

 

Peter Lindstrom 


Lawrence Livermore National Laboratory

Towards peta-scale data analysis and visualization 

With scientific data sets from observation and simulation reaching peta-
and even exabytes in the near future, our data analysis and
visualization capabilities are being outstripped by our ability to
generate such enormous volumes of data. This problem is further
exacerbated by the widening gap in processing speed afforded by
multicore architectures and virtually stagnating disk I/O performance.
To address this data explosion challenge, our research group at LLNL is
developing new scalable techniques that minimize data transfer through
compression and I/O optimal data access such as single-pass stream
processing, and via cache-oblivious data reordering for improved cache
utilization. This talk will focus on key advances in I/O efficient
processing in the context of scientific visualization, with an emphasis
on managing unstructured data such as computational meshes, isosurfaces,
graphs, and floating-point geometry and fields. 

Biography

Dr. Peter Lindstrom is a Computer Scientist in the Center for Applied
Scientific Computing at Lawrence Livermore National Laboratory. He
currently leads several research projects on analysis and visualization
of large-scale spatial data to support science applications such as
climate modeling, inertial confinement fusion, and fluid dynamics. His
recent research spans areas such as data compression, out-of-core and
cache-efficient algorithms, multiresolution modeling, terrain
visualization, mesh generation, and computer vision. Dr. Lindstrom
earned B.S. degrees in Computer Science, Mathematics, and Physics from
Elon University in 1994, received a Ph.D. in Computer Science from
Georgia Institute of Technology in 2000, and joined LLNL shortly
thereafter. 

 

 

_______________________________________________
Colloquium mailing list
[email protected]
https://secure.engr.oregonstate.edu/mailman/listinfo/colloquium

Reply via email to