IN CONSTRUCTION
This event is supported by INRIA, UIUC, NCSA, ANL, as well as by EDF
Main Topics |
Schedule |
Speaker |
Affiliation |
Type of presentation |
Title (tentative) |
Download |
|
|
|
|
|
|
|
Workshop Day 1 |
Wednesday June 13th |
|
|
|
|
|
|
|
|
|
|
TITLES ARE TEMPORARY (except if in bold font) |
|
Registration |
08:00 |
|
|
|
|
|
Welcome and Introduction |
08:30 |
Marc Snir + Franck Cappello |
INRIA&UIUC |
Background |
Welcome, Workshop objectives and organization |
|
|
08:45 |
Bertrand Braunschweig |
INRIA |
Background |
Welcome to INRIA Rennes |
|
|
09:00 |
Thierry Priol |
INRIA |
Background |
HPC @ INRIA (update) |
|
Sustained Petascale |
09:15 |
Bill Kramer |
NCSA |
Background |
Blue Waters UPDATE and performance metrics. |
|
|
09:45 |
Torsten Hoefler |
NCSA |
Background |
Blue Waters Applications and Scalability/Performance Challenges and performance modeling |
|
|
10:15 |
Break |
|
|
|
|
|
10:45 |
François Bodin |
INRIA |
Background |
OpenACC |
|
|
11:15 |
Marc Snir |
ANL |
Background |
BlueGene Q: First impression |
|
|
11:45 |
Robert Ross |
ANL |
Background |
BIG DATA |
|
|
12:15 |
Lunch |
|
|
|
|
Mini Workshop1 |
|
|
|
|
|
|
Fault tolerance |
13:30 |
Bill Kramer and Marc Snir |
UIUC, ANL |
Background |
Fault tolerance needs at NCSA and ANL |
|
|
14:00 |
Ana Gainaru |
NCSA |
Joint Result |
High precision fault prediction for Blue Waters |
|
|
14:30 |
Amina Guermouche |
INRIA |
Joint Result |
TBD |
|
|
15:00 |
Break |
|
|
|
|
|
15:30 |
Mehdi Diouri |
INRIA |
Joint Results |
Fault tolerance and energy consumption |
|
|
16:00 |
Tatiana |
INRIA |
in progress |
TBD |
|
|
16:30 |
Sanjay Kale |
UIUC |
Result |
TBD. |
|
|
17:00 |
Discussions |
|
|
How to address Petascale fault tolerance needs |
|
|
18:00 |
Adjourn |
|
|
|
|
Mini Workshop2 |
|
|
|
|
|
|
I/O and BigData |
13:30 |
Thom Dunning and Rob Ross |
UIUC, ANL |
Background |
I/O and BIGDATA needs at NCSA and ANL |
|
|
14:00 |
Gabriel Antoniu |
INRIA |
Joint Result |
TBD |
|
|
14:30 |
Matthieu Dorier |
INRIA |
Joint Result |
In-Situ Interactive Visualization of HPC Simulations with Damaris |
|
|
15:00 |
Break |
|
|
|
|
|
15:30 |
Pavan Balaji |
ANL |
Background |
Fault tolerance and energy consumption |
|
|
16:00 |
Dries Kimpe |
ANL |
Background |
TBD |
|
|
16:30 |
--- |
--- |
--- |
|
|
|
17:00 |
Discussions |
|
|
How to address Petascale I/O and Big Data needs |
|
|
18:00 |
Adjourn |
|
|
|
|
|
|
|
|
|
|
|
Workshop Day 2 |
Thursday June 14th |
|
|
|
|
|
|
|
|
|
|
|
|
Math for HPC |
08:30 |
Frederic Vivien |
INRIA |
Joint Result |
A Unified... |
|
|
09:00 |
Paul Hovland |
ANL |
Background |
TBD. |
|
|
09:30 |
Laurent Hascoet |
INRIA |
Joint Results |
Gradient of MPI-parallel codes |
|
|
10:00 |
Break |
|
|
|
|
Programming languages |
10:30 |
Rajeev Thakur |
ANL |
Background |
TBD. |
|
|
11:00 |
Sanjay Kale |
UIUC |
Background |
TBD. |
|
|
11:30 |
--- |
--- |
--- |
-- |
|
|
12:00 |
Lunch |
|
|
|
|
|
|
|
|
|
|
|
Mini Workshop3 |
|
|
|
|
|
|
Numerical libraries |
13:30 |
Paul Hovland and Bill Gropp |
UIUC, ANL |
Background |
Numerical libraries needs at NCSA and ANL |
|
|
14:00 |
Laura Grigori |
INRIA |
Joint Result |
TBD |
|
|
14:30 |
François Pelegrini |
INRIA |
Joint Result |
TBD |
|
|
15:00 |
Break |
|
|
|
|
|
15:30 |
Jocelyne Erhel |
INRIA |
Background |
TBD |
|
|
16:00 |
Daisuke Takahashi |
U. Tsukuba |
Joint Result |
TBD |
|
|
16:30 |
--- |
--- |
--- |
|
|
|
17:00 |
Discussions |
|
|
How to address Petascale Numerical Libraries needs |
|
|
18:00 |
Adjourn |
|
|
|
|
|
|
|
|
|
|
|
Mini Workshop4 |
|
|
|
|
|
|
Programing Models |
13:30 |
Rajeev Thakur and Sanjay Kale |
UIUC, ANL |
Background |
Programming model needs at NCSA and ANL |
|
|
14:00 |
Jean-François Mehaut |
INRIA |
Joint Result |
TBD |
|
|
14:30 |
Brice Goglin |
INRIA |
Background |
TBD |
|
|
15:00 |
Break |
|
|
|
|
|
15:30 |
Pavan Balaji |
ANL |
Background |
TBD |
|
|
16:00 |
Thomas Ropars |
EPFL |
Result |
Global operation optimizations on Multicore. |
|
|
16:30 |
Alexandre Duchateau |
INRIA |
Joint Result |
TBD |
|
|
17:00 |
Discussions |
|
|
How to address Petascale programing model needs |
|
|
18:00 |
Adjourn |
|
|
|
|
|
|
|
|
|
|
|
|
19:00 |
Banquet |
|
|
@ Saint Malot |
|
|
|
|
|
|
|
|
Workshop Day 3 |
Friday June 15th |
|
|
|
|
|
|
|
|
|
|
|
|
Mini Workshop5 |
|
|
|
|
|
|
Mapping and Scheduling |
08:30 |
BIll Kramer and Marc Snir |
UIUC, ANL |
Background |
Mapping and Scheduling needs at NCSA and ANL |
|
|
09:00 |
Emmanuel Jeonnot |
INRIA |
Joint Result |
TBD |
|
|
09:30 |
François Pellegrini |
INRIA |
Background |
TBD |
|
|
10:00 |
Torsten Hoefler |
NCSA --> ETH |
Background |
On-node and off-node Topology Mapping for Petascale Computers |
|
|
10:30 |
Joseph Emeras (Olivier Richard) |
INRIA |
Background |
TBD |
|
|
11:00 |
Discussions |
|
|
How to address Petascale Mapping and Scheduling needs |
|
Mini Workshop6 |
|
|
|
|
|
|
HPC/Cloud |
08:30 |
Kate Keahey and Franck Cappello |
ANL, INRIA |
Background |
An introduction to HPC Cloud |
|
|
09:00 |
Gabriel Antoniu |
INRIA |
Joint Result |
TBD |
|
|
09:30 |
Frederic Desprez |
INRIA |
Background |
TBD |
|
|
10:00 |
Bogdan Nicolae |
INRIA |
Joint Results |
TBD |
|
|
10:30 |
Derrick Kondo |
INRIA |
Result |
Characterization and Prediction of Host Load in a Google Data Center |
|
|
11:00 |
Discussions |
|
|
How to address HPC Cloud needs |
|
|
|
|
|
|
|
|
|
12:00 |
Franck Cappello and Marc Snir |
|
|
Discussion and Closing |
|
|
|
|
|
|
|
|
|
12:30 |
Lunch |
|
|
|
|
Abstracts
Laurent Hascoet: Gradient of MPI-parallel codes
Automatic Differentiation (AD) is the primary means of obtaining analytic derivatives from a numerical model given as a computer program. Therefore, it is an essential productivity tool in numerous computational science and engineering domains. Computing gradients with the adjoint mode of AD via source transformation is a particularly beneficial but also challenging use of AD. To date only ad hoc solutions for adjoint differentiation of MPI programs have been available, forcing AD users to reason about parallel communication dataflow and dependencies and manually develop adjoint communication code. In this collaboration between Argonne, RWTH Aachen, and INRIA, we characterize the principal problems of adjoining the most frequently used communication idioms. We propose solutions to cover these idioms and consider the consequences for the MPI implementation, the MPI user and MPI-aware program analysis.
Matthieu Dorier: In-Situ Interactive Visualization of HPC Simulations with Damaris
The I/O bottlenecks already present on current petascale systems force to consider new approaches to get insights from running simulations. Trying to bypass storage or drastically reducing the amount of data generated will be of outmost importance for the scales to come and, in particular, for Blue Waters.
This presentation will focus on the specific case of in-situ data analysis collocated with the simulation’s code and running on the same resources. We will first present some common visualization and analysis tools, and show the limitations of their in-situ capabilities. We then present how we enriched the Damaris I/O middleware to support analysis and visualization operations. We show that the use of Damaris on top of existing visualization packages allows us to (1) reduce code instrumentation to a minimum in existing simulations, (2) gather the capabilities of several visualization tools to offer adaptability under a unified data management interface, (3) use dedicated cores to hide the run time impact of in-situ visualization and (4) efficiently use memory through an allocation-based communication model.
Torsten Hoefler: On-node and off-node Topology Mapping for Petascale Computers
Network topology and the efficient mapping from tasks to physical processing elements is an increasing problem as we march towards larger systems. The last generation of Petascale class systems which comes right before a major switch to optical interconnects is highly affected due to their large low-bisection Torus networks. We will explore opportunities to improve communication performance by avoiding network congestion with automated task remapping. We discuss how we combine different approaches, various well-known heuristics, a heuristic based in RCM, INRIA's SCOTCH, and INRIA's tree-map algorithms for achieving highest mapping performance for on-node as well as off-node mappings. We also investigate a theoretical possibility to reduce energy usage due to minimizing dilation of the mapping. This whole work is done in the context of MPI and can readily be adapted to real production applications.
Derrick Kondo: Characterization and Prediction of Host Load in a Google Data Center
[Joint work with Sheng Di, Walfredo Cirne]
Characterization and prediction of system load is essential for optimizing its performance. We characterize and predict host load in a real-world production Google data center, using a detailed trace of over 25 million tasks across over 12,500 hosts. In our characterization, we present valuable statistics and distributions of machine’s maximum load, queue state and relative usage levels. Compared to traditional load from Grids and other HPC systems, we ?nd that Google host load exhibits higher variance due to the high frequency of small tasks. Based on this characterization, we develop and evaluate different methods of machine load prediction using techniques such as autoregression, moving averages, probabilistic methods, and noise ?lters. We ?nd that a linear weighted moving average produces accurate predictions with a 80%-95% success rate, outperforming other methods by 5%-20%. Surprisingly, this method outperforms more sophisticated hybrid prediction methods, which are effective for traditional Grid loads but not data center loads due to its more frequent and severe ?uctuations.