Multilevel Adaptive Grid Refinement and Fault Recovery


The focus of this talk will be the use of adaptive finite element multilevel refinement techniques. The benefits of such an approach will by demonstrate by briefly discussing some example applications, including data mining and plasma ion implantation. Next, we will discuss some of the implementation issues associated with working on multiple grid layers in a parallel setting. To finish the discussion, we will focus on one of our current projects, namely fault recovery.  On future extreme-scale computers, faults will become increasingly common as the number of individual components grows without a compensating improvement in reliability. One approach to manage such faults is to build the recovery process into the algorithm implementation.


Linda Stals has a strong background in the implementation of scalable algorithms on high performance computers. In the quest to simulate successively larger and more realistic models, it has become evident that such goals can not be fully achieved without addressing the design of scalable algorithms; and to realise scalable algorithms attention must be paid to both mathematical and computing ideas. Linda has a unique set of skills that allow her to build on the interface between these two disciplines.

Linda has implemented a parallel multigrid method program based on adaptive finite elements. The code developed as part of this research has formed the basis of projects designed to study applications such as: plasma ion implantation, flow through heterogeneous material (in combination with the DOUG code), radiation transport equations and thin plate splines. She has also studied cache aware multigrid methods and is currently looking at resilient iterative solvers.

In 2014, Linda was awarded a College Commendation For Outstanding Contribution to Student Learning, and nominated for a university citation; and was awarded the status as a Senior Fellow of The Higher Education Academy in 2015.

Zoom Details:

Topic: Mathematics and Computational Sciences Seminar Series 
Time: This is a recurring meeting Meet anytime

Join Zoom Meeting

Meeting ID: 880 8490 0678
Password: 971872
One tap mobile
+61731853730,,88084900678#,,,,0#,,971872# Australia
+61861193900,,88084900678#,,,,0#,,971872# Australia

Dial by your location
        +61 7 3185 3730 Australia
        +61 8 6119 3900 Australia
        +61 8 7150 1149 Australia
        +61 2 8015 6011 Australia
        +61 3 7018 2005 Australia
Meeting ID: 880 8490 0678
Password: 971872
Find your local number:

Or an H.323/SIP room system:
    Dial: +61262227588 (AUCX)
    Meeting ID: 88084900678
    H323/SIP Password: 971872

Join by Skype for Business