Agenda
December 11-13, 2006
Perseverance Hall (Building 54, Room 130)
Purpose statement
The focus of the AHM will be on identifying the data management needs of scientific application areas in the petascale regime that the SDM center should focus on in the next few years on. We hope to end up with an organization of our activities in the center around projects, where each project will be assigned technologies desired, people to do the work, and specific application people to work with.
We'll be dedicating the entire first day, Monday, Dec. 11, to hear from various application domains about their data management needs. The first half of the second day will be organized as brief presentations of technologies that the SDM center has developed and deployed over the last few years. The afternoon will be dedicated to working groups whose goal is to identify future joint projects with applications and technologies to be applied to them. The third day is planned for hearing summaries from the working groups, and forming Projects-by-Application teams.
Monday, December 11,
2006
8:00 Continental Breakfast
8:40 introductory remarks – Arie Shoshani PPT
Session 1: Targeted application domains: Fusion, Combustion)
9:00 End-to-end data management
capabilities in the GPSC & CPES Fusion: achievements and plans. Scott Klasky, ORNL
9:30 Data management challenges in combustion understanding
and modeling, Jacqueline Chen, Sandia
10:00 Free form discussion period
10:30 – 11:00 Break
Session 2: (AMR perspective with examples from: Astrophysics,
Fusion, Combustion)
11:00 Data management
requirements in combustion and astrophysics simulations using AMR techniques.
John Bell, LBNL
11:30 Data challenges in Chombo: Adaptive Mesh Refinement Library, with applications
to Fusion, Phil Colella, LBNL
12:00 Free form discussion period
12:30 – 1:30 Lunch
Session 3: (Targeted
application domains: Quantum Chemistry, Astrophysics)
1:30 Existing and anticipated
data management problems in scaling QBOX: a massively parallel code for molecular
dynamics simulations, Francois Gygi, UCD
2:00
The Spallation Neutron Source: future
data and storage requirements, Steve Miller,
Shelly Ren, ORNL
2:30 Free form discussion period
3:00 – 3:30 Break
Session 4: (Targeted application domain: Materials and
Nanostructures)
3:30 Running Multiresolution ADaptive NumErical Scientific Simulation
(MADNESS) on petascale computers: new data challenges,
Robert Harrison, ORNL
4:00 “The Petascale Data Storage
institute” - Carlos Maltzahn (for Scott Brandt), UCSC, part 1-
4:30 Free form discussion period
Session 5:
5:00 Data Management requirements for petascale computing
General discussion period and organizing into breakout groups
5:30 Session ends
Dinner on your own
Tuesday, December 12,
2006
8:00 Continental Breakfast
8:20 – organizational comments – Arie Shoshani
8:30 Very short presentations of SDM members on existing technologies
(String out 10 min presentations)
Storage Efficient Access (SEA) area
·
Rob Ross: Parallel I/O technology
· Alok Choudhary: High-Performance Parallel I/O Libraries PPT
· Ekow Otoo/Arie Shoshani: Storage Resource Management (SRM) and MPIIO-SRM PPT
· Weikuan Yu/Jeff Vetter: Testing efficiency of parallel I/O software PPT
Data Mining and Analysis (DMA) area
· Chandrika Kamath: Techniques for feature identification in scientific data PPT
· John Wu: Scientific Data Indexing, the Fastbit package PPT
· Nagiza Samatova: Automatic Parallelization for Statistical Computing with pR PPT
Scientific Workflow Automation (SPA) area
· Terence Critchlow/Ilkay Altintas: Kepler - status and plans PPT
· Mladen Vouk: experience with Astrophysics workflows and dashboard concepts PPT
· Norbert Podhorszki/Bertram Ludaescher: experience with fusion workflows PPT
10:10 – 10:30 Break
10:30 -11:30
-- “Managing exploratory workflows” - Claudio Silva, UofUtah PDF
-- “Active Storage Processing in Parallel Filesystems” - Jarek Nieplocha, PNNL PPT
11:30 – 1:30 Lunch
1:30 -3:30
Form Groups by application areas
Breakout groups start
3:30 – 4:00 Break
4:00 -5:30
Breakout groups continue
Dinner on your own
Wednesday, December
13, 2006
8:00 Continental Breakfast
8:30
Summary from breakout groups and discussion
10:00 – 10:30 Break
10:30 -12:00
Summarize requirements across applications and identify
technologies that should be applied to satisfy these requirements.
11:30 – 1:30 Lunch
1:30 -3:30
Free form discussions by areas:
SPA: Scientific Process Automation
DMA: Data Mining and Analysis
SEA: Storage Efficient Access
AHM Meeting Adjourns