• 2015 Jan - Scientific Data Management, Analysis, and Visualization PI Meeting
    01/13/2015 - 01/15/2015
    Walnut Creek
    California

2015 Jan - Scientific Data Management, Analysis, and Visualization PI Meeting

The purpose for this meeting is to provide an opportunity for project leaders to communicate with one another about their work and to foster interaction between and better awareness of activities across these different research projects. Aside from assessing the current state of research through technical talks and poster sessions, specific goals and objectives for the PI meeting are as follows:

  1. Participants will understand the hardware architecture for next-generation and future exascale computing systems and evaluate the impact the changing architecture will have on their research;
  2. Participants will understand the public version of the Exascale Preliminary Program Design Document and assess the fit of their research into that context and framework;
  3. Participants will discuss the newly-developed Data Software Stack and assess how their research maps to that diagram, identifying gaps as appropriate in cooperation with the leads for the Data Management Nexus and the Data Analytics and Visualization Sub-Plexus; and
  4. Participants will begin discussion of what will be required to integrate various research activities/projects to provide the data management and data analytics and visualization technology that are required to support scientific discovery at extreme scale.

Venue

Walnut Creek Marriott
2355 N Main St.
Walnut Creek, CA 94596
United States

Involved Programs/Projects

Lucy Nowell

Agenda

Day 1

7:00-8:30 – Breakfast and Registration

8:30-9:00 – Welcoming remarks, logistics

9:00-10:00 – Plenary talks about new projects (5 minutes each)

Ahrens, Agarwal, Bethel, Moreland, McCormick, Peterka, Samatova, Schwan, Shen

10:00-11:45 – Morning break and poster session for new projects (Ahrens, Agarwal, Bethel, Moreland, McCormick, Peterka, Samatova, Schwan, Shen)

11:45-13:30 – Working lunch

Speakers: Next-generation platform presentations, H. Wasserman (NERSC) and Al Geist (ORNL/OLCF)

13:30 – 15:00 – Exascale Plan overview presentations.

Lucy Nowell, Jim Ahrens/Wes Bethel, Rob Ross/Gary Grider

15:00-15:30 – Afternoon break

15:30-17:00 – Exascale Plan Coordination discussion groups

Two breakout/discussion groups: Data Management (DM) and Data Analytics & Visualization (DAV)

Assignment: Participants prepare one slide that states the high-level objective for their work and then identifies one or more areas in the EPPDD where their work fits. Participants complete these slides by the end of the session and email them to Wes/Jim/Rob/Gary and the project lead PI.

Project leads will present these slides in a lightning round session Day 2.

 

Day 2

7:00-8:00 – Breakfast

8:00-9:30 – Plenary technical talks: Ahrens, Bethel, McCormick, Wu (20 minutes each)

9:30-10:45 – Poster session and break: Ahrens, Bethel, McCormick, Wu

10:45-12:00 – Exascale Coordination report back and lightening talks

All participants, order TBD.

12:00-13:15 – Working lunch

Speaker: Future architecture platforms/challenges. J. Shalf (LBNL)

Assignment: Project participants each prepare one slide stating how your project(s) either help target challenges of future architectures or how your research might be affected by future architectures/challenges. Participants will present their slides in a lightning talk session on Day 3.

13:15-13:45 – Plenary data software stack overview and breakout discussion

Bethel/Ahrens, Grider/Ross.

13:45-15:30  – Breakout groups: DM and DAV. (Afternoon break available 14:45-15:30)

Assignment: Participants each prepare one slide that states the high-level objective for their work and then identifies where in the software stack their project fits/lands. Participants complete these slides by the end of the session and email them to Wes/Jim/Rob/Gary and the lead PI for the project. Participants will present this information in a lightning round session Day 2.

15:30-17:15 – Plenary: data software stack lightning talks

All participants, order TBD.

         

Day 3

7:00-8:00 – Breakfast

8:00-8:19 - Morning announcements

8:10-9:15 – Plenary technical talks (20 minutes each)

Klasky, Prabhat, Ross

9:15-10:30 – Poster session and break

Klasky, Prabhat, Ross

10:30-11:30 – Plenary: Architectural impact lightning talks and discussion

All participants, order TBD.

11:30-13:00 – Working lunch

Speaker: Program Management Expectations, L. Nowell, ASCR

13:00-14:00 – Plenary technical talks by Early Career Research Program award recipients (12 minutes each):

Childs, Kalyanaraman, Lindstrom, Rusu

14:00-14:45 – Poster session and afternoon break, ECRP projects.

Childs, Kalyanaraman, Lindstrom, Rusu

14:45-15:00 - Afternoon break

15:00-16:15– Plenary and Breakouts on Next Steps: Overview and objectives, L. Nowell

Breakouts: Participants divide into per-project groups.

Discussion topic: how do we as a community provide what is needed to result in production-quality software for the exascale platform in time?

Objectives: Each group to identify 4-6 challenges/objectives along with suggestions/ideas for meeting those challenges. The discussion leader will prepare a slide or two that summarize the challenges/ideas, and email to Lucy/Wes upon the conclusion of the discussion

16:15-17:00 – Plenary roundtable discussion: present ideas

17:00 – Adjourn

Resources

Quad Charts

Over the last two decades the Department of Energy has es- tablished the essential role of high-performance computing in advancing a wide range of...
Quad chart for PI meeting Jan 2015
Quad Chart for "Data Exploration at the Exascale" for Jan. 2015 SDMA&V PI Meeting
DataMovement Project Quad Chart for Jan. 2015 SDMA&V PI Meeting
Decaf quad chart Jan. 2015 PI Meeting
Modern computational and experimental sciences face a major challenge in coping with the sheer volume and complexity of data produced. Storing,...
The objective of this project is to develop effective data summarization and analysis techniques for simulations that run on DOE's exascale computing...
GraphKernels Project Quadchart for Jan 2015 SDMA&V PI meeting
By dynamically adapting simulation resolution across space and time, Adaptive Mesh Refinement  (AMR) simulation codes can drastically improve...
DOE Office of Science has embarked upon a revolutionary path to exascale (1018 FLOPS) supercomputing. Just as the massive computing power of these...
This project focuses on a key set of problems facing our community as we move towards the exascale regime and respond to challenges resulting from...
Our project takes aim at a set of research challenges in enabling scientific knowledge discovery within the context of in situ processing at extreme‐...
This quad chart presents the MONA project's goals and expected impact, as well as a timeline for project deliverables.
Scientific discovery at the extreme scale is a unique technical challenge, requiring the reduction of massive amounts of data into compact analysis...
OptPowerUsage quad chart for SDMA&V PI meeting, Jan 2015
RSVP Quad Chart for January 2015 Meeting
This project aims to improve scientific productivity for DOE mission critical applications with innovative data processing tools and techniques....
The Scout Domain-Specific Language  (DSL) enables scientists to productively develop mesh-based applications that are portable to different and...
SEEM Project Quad Chart for Jan 2015 SDMA&V PI Meeting
XVis Project Quad Chart for Jan 2015 SDMA&V PI Meeting

Posters

CODES Poster for January 2015 PI Meeting
Poster for Jan 2015 PI meeting
"Data Exploration at the Exascale" project poster for Jan. 2015 SDMA&V PI Meeting
One of the key challenges to achieving high performance in extreme-scale computing is how to limit data movement between the CPU, memory banks,...
Decaf poster Jan. 2015 PI Meeting
The objective of this project is to develop effective datasummarization and analysis techniques for simulations that run onDOE's exascale computing...
Modern computational and experimental sciences face a major challenge in coping with the sheer volume and complexity of data produced. Storing,...
GraphKernels Project Poster for Jan. 2015 SDMA&V PI Meeting
By dynamically adapting simulation resolution across space and time, Adaptive Mesh Refinement  (AMR) simulation codes can drastically improve...
DOE Office of Science has embarked upon a revolutionary path to exascale (1018 FLOPS) supercomputing. Just as the massive computing power of these...
This project focuses on a key set of problems facing our community as we move towards the exascale regime and respond to challenges resulting from...
Our project takes aim at a set of research challenges in enabling scientific knowledge discovery within the context of in situ processing at extreme‐...
This poster presents the different elements of the MONA project, starting with the idea of online monitoring combined with monitoring data analytics...
Scientific discovery at the extreme scale is a unique technical challenge, requiring the reduction of massive amounts of data into compact analysis...
OptPowerUsage poster for SDMA&V meeting Jan 2015
   The advent of the exascale era is forcing a re-examination of the computational processes for science, and particularly those for scientific data...
This project aims to improve scientific productivity for DOE mission critical applications with innovative data processing tools and techniques....
The Scout Domain-Specific Language (DSL) enables scientists to productively develop mesh-based applications that are portable to different and future...
SEEM Project Poster for Jan 2015 SDMA&V PI Meeting
Amid today’s challenges of constrained power budgets and rapidly changing processor and system architecture designs, this project aims to study a...
XVis Project Poster for Jan 2015 SDMA&V PI Meeting.

Handouts

Over the last two decades the Department of Energy has es- tablished the essential role of high-performance computing in advancing a wide range of...
The goal of the CODES project is use highly parallel simulation to explore the design of exascalestorage architectures and distributed data-intensive...
Handout for PI meeting Jan 2015
DataExplor Project Handout for Jan. 2015 SDMA&V PI Meeting
This project is developing new techniques for reducing data movement in extreme-scale computing by improving data and compute locality, and by...
Decaf handout Jan. 2015 PI Meeting
The objective of this project is to develop effective datasummarization and analysis techniques for simulations that run onDOE's exascale computing...
The challenges of productively programming the exascale generation of high-performance computing systems has led to a groundswell of research...
Modern computational and experimental sciences face a major challenge in coping with the sheer volume and complexity of data produced. Storing,...
This document is a PDF version of the two-page handout showing (1) the major focus/project areas for DM and DAV in the exascale plan, and (2) a draft...
GraphKernels Project One Page Handout for Jan 2015 SDMA&V PI meeting
By dynamically adapting simulation resolution across space and time, Adaptive Mesh Refinement  (AMR) simulation codes can drastically improve...
DOE Office of Science has embarked upon a revolutionary path to exascale (1018 FLOPS) supercomputing. Just as the massive computing power of these...
This project focuses on a key set of problems facing our community as we move towards the exascale regime and respond to challenges resulting from...
Our project takes aim at a set of research challenges in enabling scientific knowledge discovery within the context of in situ processing at extreme‐...
The MONA project seeks to instrument the data management workflows associated with exascale applications. Instrumentation serves the purpose of...
The MONA project seeks to instrument the data management workflows -- online analytics and visualization -- present on future exascale machines. The...
Scientific discovery at the extreme scale is a unique technical challenge, requiring the reduction of massive amounts of data into compact analysis...
OptPowerUsage handout for Jan. 2015 SDMA&V Meeting.
The advent of the exascale era is forcing a re-examination of the computational processes for science, and particularly those for scientific data...
This project aims to improve scientific productivity for DOE mission critical applications with innovative data processing tools and techniques....
SEEM Pproject Handout for Jan 2015 SDMA&V PI Meeting
XVis Project One Page Handout for Jan 2015 SDMA&V PI Meeting.

Presentations

Amid today’s challenges of constrained power budgets and rapidly changing processor and system architecture designs, this project aims to study a...
J. Shalf lunchtime presentation 1/14/2015 at the January 2015 SMDAV PI meeting
Presentation from the PI meeting
Presentation for "Data Exploration at the Exascale" for Jan. 2015 SDMA&V PI Meeting.
Decaf presentation Jan. 2015 PI Meeting
PDF of the slide deck presented by Al Geist (ORNL) during lunch on 1/13/2015 at the SDMAV program wide meeting
GraphKernels Project Presentation for Jan. 2015 SDMA&V PI meeting
HPVA project presentation
DOE Office of Science has embarked upon a revolutionary path to exascale (1018 FLOPS) supercomputing. Just as the massive computing power of these...
Scientific discovery at the extreme scale is a unique technical challenge, requiring the reduction of massive amounts of data into compact analysis...
Runtime System for i/o staging in support of Voluminous in-situ Processing of extreme scale data (RSVP) presentation for Jan. 2015 PI meeting
The challenges of productively programming the exascale generation of high-performance computing systems has led to a groundswell of research...
SEEM Project Presentation for Jan 2015 SDMA&V PI Meeting
XVis Project Presentation for January 2015 SDMA&V PI Meeting