The Computational Science and Engineering Petascale Initiative at LBNL identified key application areas with specific needs for advanced programming models, algorithms, and other support. The applications were chosen to be consistent with the mission of the Department of Energy, with a particular focus on applications that benefit energy research, those supported by other Recovery Act funding and Energy Frontier Research Centers (EFRCs). The initiative paired post-doctoral researchers with these high-impact projects at NERSC. The post-doctoral research program was centered at the Oakland Scientific Facility (OSF), NERSC's operations center and was led by principal investigator Alice Koniges.
The multicore revolution provides opportunity for major increases in computational power over the next several years, if it can be harnessed. Elective use of multicore platforms is essential for science problems that require petascale computing, but is also necessary to continue the performance growth associated with Moore's Law at computing systems at all scales. This transition from simply increasing the single-processor and network performance to a different architectural paradigms forces application programmers to rethink the basic models of algorithm development and parallel programming from both the language and problem division standpoints. Additionally, advances in architecture and availability of resources are making the high performance computing (HPC) world more attractive to a variety of research areas that were not previously using advanced computing.
- Two post-docs (Kim and Preissl) were co-authors and significant contributors to the paper “Application Acceleration on Current and Future Cray Platforms,” winner BEST PAPER, CUG 2010, Edinburgh, May 2010.
- Two first author (Fagnan and Narayanan) post-doc papers were presented at CUG 2011.
- More than 20 refereed publications, talks, and conference poster presentations have been given.
- Two post-docs (Kim and Maia) served as teaching assistants for the NERSC sponsored user course on GPUs, Aug. 2010.
- One post-doc (Preissl) served as a teaching assistant for SC10,11 (Supercomputing) conference tutorial on PGAS Languages
- One post-doc (Maia) created and launched a scientific data bank for x-ray images: Coherent X-ray Imaging Data Bank.
- One post-doc produced material for mini-apps for SC12 tutorials
- Several new refereed publications appearing in 2013 and 2014
Where Are They Now?
|Name||NERSC Postdoc Location and Project||Position After Postdoc|
|Brian Austin||LBNL's Center for Beam Physics, simulating next generation light sources based on x-ray free electron lasers||NERSC Advanced Technology Group|
|Kjiersten Fagnan||Collaboration with the Center for Computational Science and Engineering group at LBNL, worked on numerical simulation of carbon sequestration and porous media flow.||NERSC User Services Group|
|Christos Kavouklis||Collaboration with LBNL's Applied Numerical Algorithms Group, working on the fast computation of volume potentials on structured grids using the method of local corrections.||LBNL, Computational Research Division|
|Jihan Kim||DOE Energy Frontier Research Center for gas separations relevant to clean energy technologies, developed a GPU code to accelerate screening of microporous materials to absorb carbon dioxide gas from power plant flue gases.||Assistant Professor, Chemical & Biomolecular Engineering, Korea Advanced Institute of Science and Technology|
|Bobby Liu||DOE Heavy Ion Sicence Virtual National Laboratory and LBNL's NDCX-II accelerator, which was also initially ARRA funded, worked to improve the ALE-AMR code and develop scalable solutions for new physics models at NERSC.|
|Filipe Maia||LBNL Advanced Light Source, accelerating biological imaging codes and Earth Sciences Division, accelerating geophysical imaging codes with GPUs||LBNL followed by Molecular Biophysics Laboratory, Uppsala University|
|Praveen Narayanan||Worked on performance characterization and benchmarking of several parallel applications running at NERSC, collaborating with code teams at several DOE facilities||nVIDIA|
|Robert Preissl||Worked with physicists at the Princeton Plasma Physics Laboratory (PPPL) to implement PGAS and hybrid programming solutions for highly scalable particle-in-cell codes||IBM|
|Xuefei (Rebecca) Yuan||LBNL Computational Research Division, improved a hybrid linear software package based on the Schur complement method||Bank of America|