Welcome at the Massively Parallel Systems Group!

The Massively Parallel Systems Group (MPS) is dedicated to researching and teaching computer architecture, with a specific focus on massively parallel systems. These days massively parallel systems are present everywhere – in our smartphones to supercomputers. They help us to do many things which were not possible before. For example, the recent stupendous success of machine learning, especially deep learning, is mainly due to the exponential increase in computational power. As such, machine learning is not new. Machine learning methods are around since the 1950s. What has predominately changed now is the computing power, with many-core processors such as GPUs as the main drivers. If these massively parallel processors are not utilized properly, they are very expensive in terms of power and energy consumption, which is not good as we aspire to reduce our carbon footprint. At MPS, we are developing new techniques to make computing devices more performance and energy-efficient while also improving their programmability. Interested in collaborating with us? Contact us, join us!

News

2025-03-19: Indo-German Collaborative Project Reccomended for Funding

We are happy to announce that our research proposal, titled "MELODY: Machine Learning and Remote Sensing Synergy for Sustainable Crop Production", has been recommended for funding by the Indo-German Science & Technology Centre (IGSTC).

This truly interdisciplinary initiative combines expertise in remote sensing, machine learning, HPC, and agriculture to address key challenges in sustainable crop production. The project brings together leading academic and industry partners from Germany and India, including Spacenus GmbH, IIT Ropar, and AgriMatrix India Pvt. Ltd.

The project has a budget of approximately 1.1 million Euros. Prof. Sohan Lal serves as a Principal Investigator and the Project Coordinator.

A heartfelt thank you to TUHH for its support and to the funding agency for making this initiative possible! We look forward to driving innovation in sustainable agriculture.

2025-03-17: MPS Paper Accepted at Computing Frontiers 2025

The paper titled "Enhancing Practicality of Memory Compression for GPUs with High-Throughput Simplifications" by Manuel Renz and Sohan Lal has been accepted for publication at the 22nd ACM International Conference on Computing Frontiers (CF) 2025. Congratulations to Manuel for the great work!

2025-01-11: TUHH and NXP Collaborative Paper Accepted at LATS 2025

The paper titled "Case Study: AI-Driven Log Extraction and Trace Outlier Detection for Efficient Post-Silicon Validation" by Kowshic Ahmed Akash, Tobias Wulf, Torsten Valentin, Alexander Geist, Ulf Kulau, John Jose and Sohan Lal has been accepted for publication at the 26th IEEE Latin American Test Symposium (LATS) 2025. This is a collaborative effort between Massively Parallel Systems and Smart Sensors groups at TUHH and NXP Semiconductors Hamburg. Congratulations to the team!

2025-01-07: Best Poster Award at VLSID 2025

The paper titled "AI-Driven Anomaly Detection in Oscilloscope Images for Post-Silicon Validation" by Kowshic A. Akash, Tobias Wulf, Torsten Valentin, Alexander Geist, Ulf Kulau and Sohan Lal has won the Best Poster Award at the 38th International Conference on VLSI Design (VLSID) 2025. Congratulations to the team on this outstanding achievement!

2024-10-20: TUHH and NXP Collaborative Paper Accepted at VLSID 2025

The paper titled "AI-Driven Anomaly Detection in Oscilloscope Images for Post-Silicon Validation" by Kowshic Ahmed Akash, Tobias Wulf, Torsten Valentin, Alexander Geist, Ulf Kulau and Sohan Lal has been accepted for publication at the 38th International Conference on VLSI Design (VLSID) 2025. This is a collaborative effort between Massively Parallel Systems and Smart Sensors groups at TUHH and NXP Semiconductors Hamburg. Congratulations to the team!(Preprint)

2024-10-07: Prof. Lal Appointed Track Chair for ISVLSI 2025

Prof. Lal has been invited to serve as a track chair for the "System Design and Security" (SDS) track at the upcoming IEEE International Symposium on VLSI (ISVLSI 2025), to be held in Greece.

2024-10-30: DFG Reccomends Funding

The research proposal titled "LEAP: Locality-Driven High-Performance and Energy-Efficient GPUs in the Post-Dennard Era" has been recommended for funding by German Research Foundation (Deutsche Forschungsgemeinschaft). The proposal focuses on optimizing data locality—reducing unnecessary data movement, which is a major energy bottleneck in modern computing. Improving data locality can significantly enhance the performance and energy efficiency of GPUs. The importance of this research is reinforced by NVIDIA’s introduction of Thread Block Clusters in Hopper and Blackwell GPUs, which also aim to access data closer to processing cores. The project has a budget of approximately 282K Euros and Prof. Lal is the Principal Investigator. Thank you to TUHH for the support and the funding agency for making this proposal a reality!

2024-09-27: Best Paper Award

The paper titled "Benchmarking Thread Block Cluster" by Tim Lühnen, Tobias Marschner, and Sohan Lal has won the Best Paper Award at 28th Annual IEEE High Performance Extreme Computing Conference (HPEC) 2024. Congratulations to the team on this outstanding achievement! (Preprint)

2024-08-29: Collaborative Paper from TUHH, IIT Ropar, Spacenus GmbH, & University of Ottawa Accepted at EJA

The paper titled "SSGAN: Cloud Removal in Satellite Images using Spatiospectral Generative Adversarial Network" by Sushil Ghildiyal, Neeraj Goel, Simrandeep Singh, Sohan Lal, Riazuddin Kawsar, Abdulmotaleb El Saddik, Mukesh Saini has been accepted for Publication at a Special Issue of European Journal of Agronomy (EJA) 2024. This is a collaborative effort by IIT Ropar, TUHH, Spacenus GmbH, and University of Ottawa. Congratulations to the team!

2024-08-27: MPS Paper Accepted at HPEC 2024

The paper titled "Benchmarking Thread Block Cluster" by Tim Lühnen, Tobias Marschner, and Sohan Lal has been accepted for Publication at 28th Annual IEEE High Performance Extreme Computing Conference (HPEC) 2024. The paper has received an Outstanding Paper Award. Congratulations to the team and especially Tim for his first Ph.D. paper! (Preprint)

2024-07-08: Prof. Lal Invited to Join Program Committee of DATE 2025

Prof. Lal has been invited to serve as a technical program committee member of the 28th edition of Design, Automation and Test in Europe Conference (DATE), which will be held from 31 March to 02 April, 2024 in Lyon, France. More details can be found at (DATE). Please submit your best work!

2024-07-04: Prof. Lal was Invited for an Interactive Session at CGI Hamburg

Prof. Lal was invited to an interactive session focused on navigating academic and cultural avenues in Germany for Indian students in the Hamburg region. It was a fantastic experience to hear the students' perspectives and share my own academic journey in Germany. Special thanks to the Consulate General of India in Hamburg (CGI, Hamburg) for the invitation!

2024-03-13: Prof. Lal invited to join a Ph.D. Defense Committee@UNISA

Prof. Lal has been invited to join the Ph.D. defense committee of Mr. Majid Salimibeni at the University of Salerno, Italy.

2024-03-05: MPS at DAC 2024

The manuscript "Pixels to Precision: Harnessing AI for Anomaly Detection in Post-Silicon Validation" by Kowshic Ahmed Akash, Tobias Wulf, Torsten Valentin, Alexander Geist, Ulf Kulau and Sohan Lal has been accepted as work in progress (WIP) at the International Design Automation Conference (DAC) 2024. Congratulations to Kowshic and the team!

2024-03-01: MPS at HPCA 2024

Tim Lühnen is attending the International Symposium on High-Performance Computer Architecture (HPCA) conference being held in Edinburgh from March 2nd to March 6th.

2024-02-27: MPS Research Poster Accepted at ISC 2024

The research poster "Auto-TBC: Auto-Optimizing CUDA Applications Using Thread Block Clusters" by Tim Lühnen, Alexander Brosig, and Sohan Lal has been accepted for presentation at the International Supercomputing Conference (ISC) 2024. Congratulations to Tim and Alex!

2024-02-07: ECIUn+ Awarded 5000 € Grant for GPU Micro Module

The module "GPU Architectures and Programming" has been accepted as a micro module for ECIU students under the ECIUn+ project. Prof. Lal has been granted 5000 € for the same. The micromodule will be offered in the SoSe 2024 and SoSe 2025. Excited to Host ECIU University Students!

2023-12-14: Prof. Lal was on a Ph.D. Defense Committee@TUB

Prof. Lal was on the Ph.D. defense committee of Kaijie Fan at Technical University of Berlin. Ms Fan successfully defended her Ph.D. thesis titled "Predictive Modeling and Optimization for Energy-Efficient GPU Computing". Congratulations to Dr. Fan!

2023-09-27: Prof. Lal Delivered an Invited Talk

Prof. Lal delivered an invited talk titled "Harnessing the Power of GPUs: A Path to Efficiency and Excellence" at the chair of computational mathematics at TUHH. We appreciate the valuable discussions that ensued!

2023-09-19: Three PhD Scholars from IIT Jammu Visited MPS

Three PhD scholars from IIT Jammu (Anjum Riaz, Gaurav Kumar, and Mahendra Kumar Gurve) visited our group. Thanks for the insightful discussions!

2023-09-12: Prof. Lal Invited to Join Program Committee of DATE 2024

Prof. Lal has been invited to serve as a technical program committee member of the 27th edition of Design, Automation and Test in Europe Conference (DATE), which will be held from 25 March to 27 March, 2024 in Valencia, Spain. More details can be found at (DATE). Please submit your best work!

2023-08-25: MPS Paper Accepted at ICCD 2023

The paper "Beyond Compression Ratio: A Throughput Analysis of Memory Compression Techniques for GPUs" by Manuel Renz and Sohan Lal has been accepted for publication at the 41st IEEE International Conference on Computer Design (ICCD) 2023. Congratulations to the team!

2022-12-19: Prof. Lal Invited to Join Program Committee of Euro-Par 2023

Prof. Lal has been invited to serve as a technical program committee member of the 29th International European Conference on Parallel and Distributed Computing (Euro-Par 2023), which will be held from 28 August to 1 September, 2023 in Cyprus. Euro-Par is the prime European conference covering all aspects of parallel and distributed processing. Abstract submission deadline is on February 17, 2023 and paper submission deadline is on February 24, 2023. More details can be found at (Euro-Par). Please submit your best work!

2022-11-16: MPS is Growing!

Mr. Tim Lühnen joined the MPS group as a research assistant/Ph.D. student. Welcome, Tim! Tim will be strengthing the teaching and research of the MPS group. Tim is interested in GPU architectures as well as RISC-V processors. We are excited to have him on the team and looking forward to expanding the research portfolio of MPS.

2022-07-01: Manuel Renz Joins MPS as a Ph.D. student

Mr. Manuel Renz joined the MPS group as a Ph.D. student. Welcome, Manuel! Manuel is interested in computer architecture and he will be working to improve the architecture of multi-/many-core processors, in particular the memory system. We are excited to explore new research opportunities with him.

2022-02-21: MPS Paper Accepted at IJPP 2022

The paper "A Quantitative Study of Locality in GPU Caches for Memory-Divergent Workloads" by Sohan Lal, Sharatchandra V. Bogaraju, and Ben Juurlink has been accepted for publication at the International Journal of Parallel Programming (IJPP) 2022. Congratulations to the team!

2022-01-22: MPS Paper Accepted at IPDPS 2022

The paper "Memory Access Granularity Aware Lossless Compression for GPUs" by Sohan Lal, Manuel Renz, Julian Hartmer, and Ben Juurlink has been accepted for publication at the International Conference on Parallel and Distributed Systems (IPDPS) 2022. Congratulations to the team!

2021-09-01: Sohan Lal Joins TUHH

Dr. Sohan Lal joins TU Hamburg as a Junior Professor to lead the massively parallel systems group.