Ioannis Vardas
Research in High Performance Computing
I am a Pre-Doctoral researcher in the TU Wien at the Research Group Parallel Computing. My research interests revolve around the High Performance Computing (HPC) area. Specifically, performance optimization of HPC applications and MPI libraries such as Open MPI and MPICH. My past research includes topics on Interconnects, Resilience and Fault-tolerance.
I developed a new MPI profiling tool called mpisee that follows a new communicator-centric approach to profile MPI applications. In this way, we can explore the use of virtual topology and the impact of MPI communicator mapping on application performance. My current research thesis is focused on the mapping and colocation of parallel applications to deep resource hierarchies that have become more common of HPC systems.
Publications
- Ioannis Vardas, Sascha Hunold, Philippe Swartvagher and Jesper Larsson Träff: Exploring Mapping Strategies for Co-allocated HPC Applications, PhD Symposium of Euro-Par 2023
- Ioannis Vardas, Sascha Hunold, Jordy I. Ajanohoun and Jesper Larsson Träff: mpisee: MPI Profiling for Communication and Communicator Structure, presented at the HIPS Workshop of IPDPS 2022
- Ioannis Vardas, Manolis Ploumidis and Manolis Marazakis: Exploring the impact of node failures on the resource allocation for parallel jobs, 14th Workshop on Resiliency in High Performance Computing (Resilience) in Clusters, Clouds, and Grids
- Ioannis Vardas, Manolis Ploumidis and Manolis Marazakis: Towards Communication Profile, Topology and Node Failure Aware Process Placement, 2020 IEEE 32nd International Symposium on Computer Architecture and High Performance Computing
- Philippe Swartvagher, Sascha Hunold, Jesper Larsson Träff and Ioannis Vardas: Using Mixed-Radix Decomposition to Enumerate Computational Resources of Deeply Hierarchical Architectures, presented at the ExaMPI Workshop of SC 2023
- Jesper Larsson Träff and Ioannis Vardas: Library Development with MPI: Attributes, Request Objects, Group Communicator Creation, Local Reductions, and Datatypes, EuroMPI '23: Proceedings of the 30th European MPI Users' Group Meeting
- Jesper Larsson Träff, Sascha Hunold, Ioannis Vardas and Nikolaus Manes Funk: Uniform Algorithms for Reduce-scatter and (most) other Collectives for MPI, 2023 IEEE International Conference on Cluster Computing (CLUSTER)
- Hunold, Sascha and Ajanohoun, Jordy I. and Vardas, Ioannis and Träff, Jesper Larsson: An Overhead Analysis of MPI Profiling and Tracing Tools, PERMAVOST '22: Workshop on Performance EngineeRing, Modelling, Analysis, and VisualizatiOn Strategy
- Dimitris Giannopoulos, Nikos Chrysos, Evangelos Mageiropoulos, Giannis Vardas, Leandros Tzanakis, Manolis Katevenis: Accurate Congestion Control for RDMA Transfers, 2018 IEEE/ACM International Symposium on Networks-on-Chip (NOCS)
Projects
- FWF-Project Algorithm Engineering for Process Mapping
- European Horizon 2020 project ExaNeSt - European Exascale System Interconnect and Storage
- Green Computing Node for European micro-servers project Euroserver
Side projects
Other interests
Research
Development tools
- Emacs and Org mode