Ongeveer 17 uur geleden - Erasmus University Rotterdam (EUR)
Rotterdam School of Management (RSM), Erasmus University, invites applications for a post-doctoral research position in the area of behavioural operations. The…
Are you interested in developing ground-breaking techniques to verify the correctness of software, and do you think parallel computing is exciting? In the GEARS project, you can do research on how to use graphics processors ...
The Model Driven Software Engineering section at Eindhoven University of Technology (TU/e) offers one PhD position on the development of GPU accelerated symbolic model checking techniques, within the NWO TOP project:
GEARS: GPU Enabled Accelerated Reasoning about System designs
The development of complex hard- and software systems is error-prone and costly. One technique that can provide very valuable feedback on the correctness of system designs is model checking. It involves exhaustively analysing a system design to determine whether it satisfies desirable functional properties. However, it is computationally very demanding.
The GEARS project is led by dr.ing. Anton Wijs. At the TU/e, he develops model checking techniques that are accelerated by the use of graphics processing units (GPUs). GPUs offer great potential for parallel computation, while keeping power consumption low. However, not all types of computation can trivially be performed on GPUs, sometimes the algorithms need to be entirely redesigned. Currently, he has an explicit-state model checker, called GPUexplore, that can run hundreds of times faster than conventional model checking approaches.
Regarding symbolic model checking techniques, IC3 is currently state-of-the-art; it can analyse large designs in reasonable time, but its limitations still mean that it is not used very commonly for the design of systems today. IC3 uses SAT solving, or the solving of Boolean or propositional satisfiability problems, as a subroutine.
In the GEARS project, the candidate will investigate whether GPUs can be employed for IC3. First of all, the candidate will investigate how SAT solving can be accelerated using GPUs, specifically focussing on SAT problems that stem from model checking. Second of all, (s)he will investigate how IC3 can benefit from using GPUs. IC3 is reportedly parallelisable, but how GPUs can be used to do so most effectively is still an open question.
The project potentially has an enormous impact, since IC3 is currently state-of-the-art in symbolic model checking, and SAT solving is used not only for model checking, but in many different application domains.
The research will be conducted in the Model Driven Software Engineering section of the Computer Science department at TU/e. This section consists of two research groups, Formal System Analysis (FSA) and Software Engineering and Technology (SET). The FSA group studies techniques to model and analyse discrete system behaviour in a mathematically rigorous way, while the SET group develops methods and tools for time- and cost-efficient evolution of high-quality software systems.
We are looking for candidates that meet the following requirements:
Middels onderstaande knop kun je direct solliciteren op deze vacature.Reageer op deze vacature