SC13 Home > SC13 Schedule > SC13 Presentation -

SCHEDULE: NOV 16-22, 2013

When viewing the Technical Program schedule, on the far righthand side is a column labeled "PLANNER." Use this planner to build your own schedule. Once you select an event and want to add it to your personal schedule, just click on the calendar icon of your choice (outlook calendar, ical calendar or google calendar) and that event will be stored there. As you select events in this manner, you will have your own schedule to guide you through the week.

SESSION: Mystic: a framework for optimal performance modeling and risk

EVENT TYPE: Emerging Technologies

TIME: 4:10PM - 4:30PM

Presenter(s):Michael McKerns

ROOM:Booth 3947

From a broad perspective, our ongoing work in the mystic project concerns the development and implementation of a comprehensive, rigorous and algorithmic framework for uncertainty quantification (UQ), thus enabling the rigorous calculation of critical system design elements (such as value at risk and hedge) on systems operating under uncertainty. Our mathematical and software infrastructure supports Caltechs Predictive Science Academic Alliance Program Center (PSAAP) and the Exascale Co-design Center for Materials in Extreme Environments (ExMatEx). Our framework is built to address model uncertainty and optimal design questions such as: How good is my model? How can I best improve my model? Given the uncertainty on the system performance, does it make sense to push certain design boundaries? Is there an optimal representation of the system which reliably predicts performance? How can I find it? Can I turn the problem of finding the best model given computational constraints and available information into an algorithm? Questions like these become especially relevant in the design of the first generation of exascale computing platforms, where it is expected the systems will operate under high rates of failure. We will discuss ongoing work to extend mystic to the computation of optimal statistical estimators, and how this may be leveraged in the optimal design of future exascale systems.

Chair/Presenter Details:

Michael McKerns - California Institute of Technology

Add to iCal  Click here to download .ics calendar file

Add to Outlook  Click here to download .vcs calendar file

Add to Google Calendarss  Click here to add event to your Google Calendar