PerMedCoE webinars are open to everyone interested in PerMedCoE tools and activities. The webinars will include a 30-40 minutes presentation and a Q&A section of around 15 minutes. The recording of this webinar will be publicly available on this web page and the PerMedCoE YouTube channel.
|Speakers:||Adam Šmelko (Charles University in Prague, Czech Republic)|
|Date and time:||Thursday 15 June 2023, 15-16 CEST|
Biomedical software developers
Developers of scientific software
|Learning outcomes:||Describe GPU threads architecture|
Find suitable parallelization opportunities
Define the first steps in coding on GPUs
|Watch the recording|
|Access the slides|
Nowadays, GPU programming plays an important role in software development. The massive parallel hardware enables us to analyse bigger datasets, process data in finer details or visualise simulations in real-time. Still, due to the programming complexity and the lack of knowledge, GPUs are not regularly used in many scientific fields. Thankfully, GPU programming matures and provides simpler tools to utilise GPU hardware. In this presentation, we describe multiple GPU programming frameworks (ranging from high-level to low-level) and showcase them on selected bioinformatics algorithms.
About the speaker:
Adam Šmelko is a third year doctoral student in the Department of Distributed and Dependable Systems at Charles University in Prague with the dissertation topic of Employing parallel computing in data-intensive tasks.
He specialises in GPU programming and in the analysis of memory accesses in programs. His current research topic involves development of tools for effective memory layouts and traversals of data structures.