
M.Sc. Urvij Saroliya
Doctoral Candidate
Technical University of Munich
School of Computation, Information and Technology
Department of Computer Engineering
Informatics 10 – Chair of Computer Architecture and Parallel Systems (Prof. Dr. Martin Schulz)
Postal Address: Boltzmannstraße 3, 85748 Garching bei München
Room: 01.04.057
Tel: +49 (89) 289 - 17680
Email: urvij.saroliya(at)tum.de
Research Interests
- High Performance Computing (HPC)
- Performance Analysis and Modeling for Distributed GPU Applications
- Heterogeneous Architectures and Accelerators
- AI/ML for HPC Resource Management
- Power/Energy Aware HPC
- HPC-AI Integrations
If you are interested in any of the above mentioned topics, feel free to contact me for BA/MA/GR.
Currently, I offer the following topic as GR or IDP: Accelerating Scientific ML through Multi-GPU Scaling
Projects
Publications
- Urvij Saroliya, Eishi Arima, Martin Kronbichler “Unified Performance Modeling Stack for Distributed GPU Applications: Complementing Analytical Insights with Machine Learning” In ACM SRC at The International Conference for High Performance Computing, Networking, Storage, and Analysis (SC Poster), November 2025. [To appear]
- Stepan Vanecek, Manuel Walter Mußbacher, Dominik Größler, Urvij Saroliya, and Martin Schulz “MT4G: A Tool for Reliable Auto-Discovery of NVIDIA and AMD GPU Compute and Memory Topologies” In Proceedings of ACM/IEEE SC'25 Workshops of The International Conference on High Performance Computing, Network, Storage, and Analysis (SC-W), pp.xx--xx, November 2025. [To appear]
- Urvij Saroliya, Eishi Arima, Dai Liu, Martin Schulz "Reinforcement Learning-driven Co-scheduling and Diverse Resource Assignments on NUMA Systems" In Proceedings of IEEE International Conference on Computer Design (ICCD), pp.170-178, November 2024.
- Urvij Saroliya “Reinforcement Learning based Resource Management for HPC Systems” Master’s Thesis in Informatics at Technical University of Munich (TUM), December 2023
- Urvij Saroliya, Eishi Arima, Dai Liu, Martin Schulz "Hierarchical Resource Partitioning on Modern GPUs: A Reinforcement Learning Approach" In Proceedings of IEEE International Conference on Cluster Computing (CLUSTER), pp. 185-196, November 2023.
- Andreas Koch, Gabriel Dax, Michael Petry, Harvey Gomez, Amir Raoofy, Urvij Saroliya, Max Ghiglione, Gianluca Furano, Martin Werner, Carsten Trinitis, Martin Langer "Reference Implementations for Machine Learning Application Benchmark" In 2023 European Data Handling & Data Processing Conference (EDHPC), pp. 1-3. October 2023.