Stefano Markidis

Hi! My name is Stefano ([ˈsteːfano]) and I was born and raised in an Italian-Greek family in Parma, a lovely city in the North of Italy, famous for the food (the delicious Parmesan cheese, Parma ham and also Barilla's pasta, and Mutti's tomato sauce) and used-to-be-good football team (some of the football aficionados might remember Tomas Brolin who is our idol in Parma). Since 2021, I am also a proud Swedish citizen.

To people asking why to pursue a career in research and teaching, I would say that higher education and research are my airplane (to rephrase one of the Red Hot Chili Peppers songs): I got to know and love the most interesting people I could never meet if I decided to stay in Parma and I saw many breath-taking places all the around the world. After high school, I moved to Torino and studied Nuclear Engineering at the Politecnico di Torino where I graduated in 2002. I got a once-in-lifetime chance of learning how to do research, at the Los Alamos National Laboratory, thanks to Prof. Giovanni Lapenta. After four years in the enchanted land, known also as New Mexico, I moved to Illinois to pursue a Ph.D. in Nuclear Engineering at the University of Illinois at Urbana-Champaign under the supervision of Prof. Rizwan-Uddin. I did the last year of my Ph.D. as an intern at the Lawrence Berkeley National Laboratory in the beautiful bay area. I obtained the Ph.D. in 2010 and moved back to Europe to work at KULeuven in Belgium. From 2012, I work in Sweden at the KTH: I started first at the PDC under the supervision of Prof. Erwin Laure and became Assistant Professor in High-Performance Computing in 2014 at CST. I became Associate Professor in High-Performance Computing in 2017. Since 2018, I am a docent in Computer Science with a specialization in High-Performance Computing. From April to June 2021, I am visiting Professor at Politecnico di Milano in Italy.

At KTH, I am a researcher, leader of the KTH HPC group, teacher, and project manager.

Research

I have been always fascinated by the idea of using the most advanced technologies, such as supercomputers, to model physical reality by means of extreme-scale simulations.

At KTH, I lead the KTH HPC group.

Our research is at the intersection of programming models, computer architectures (GPUs, FPGAs, and heterogeneous memory and storage systems), scientific computing (computation fluid dynamics, radiation treatment planning, and computational plasma physics), and deep-learning on the road to the next supercomputing milestone called the exascale era (supercomputer that will deliver exa = 10^18 floating-point operations per second on the HPC benchmark called HPL).

Some of our recent papers:

• Markidis, S. Physics-Informed Deep-Learning for Scientific Computing, submitted to Frontiers.
• Karp M, Podobas A, Jansson N, Kenter T, Plessl C, Schlatter P, Markidis S. High-Performance Spectral Element Methods on Field-Programmable Gate Arrays. Accepted at IPDPS 2021.
• Chien SW, Podobas A, Peng IB, Markidis S. tf-Darshan: Understanding Fine-grained I/O Performance in Machine Learning Workloads.  Cluster 2020
• Markidis S, Peng I, Podobas A, Jongsuebchoke I, Bengtsson G, Herman P. Automatic Particle Trajectory Classification in Plasma Simulations. AI4S Workshop at SC20.
• Markidis S, Olshevsky V, Toth G, Chen Y, Peng I, Lapenta G, Gombosi T. Kinetic Modeling of Magnetospheres. AGU Book. 
• Karp M, Jansson N, Podobas A, Schlatter P, Markidis S. Optimization of tensor-product operations in Nekbone on GPUs. Poster at SC20. 

I am also known as the main architect of the iPIC3D code, a C++ implicit Particle-in-Cell (PIC) code for space physics simulations, and an inventor of new PIC numerical methods, such as:

• Markidis S, Olshevsky V, Sishtla CP, Chien SW, Laure E, Lapenta G. PolyPIC: the Polymorphic Particle-in-Cell Method for Fluid-Kinetic Coupling. Frontiers in Physics. 2018 
• Markidis S, Henri P, Lapenta G, Rönnmark K, Hamrin M, Meliani Z, Laure E. The fluid-kinetic particle-in-cell method for plasma simulations. Journal of Computational Physics. 2014 
• Markidis S, Lapenta G. The Energy Conserving Particle-in-Cell Method. Journal of Computational Physics. 2011.

In 2005 and 2017, I was a recipient of the prestigious R&D100 award.

We are also the recipient of the best paper awards at AsHES, EuroMPI, HiPC, and IPDPS.

Teaching

At KTH, I am responsible for the High-Performance Computing sub-track in the Computer Science program. 

I am course responsible, teacher, and examiner for the main courses of the HPC track:

• DD2358 - Introduction to High-Performance Computing - P3. Topics: Computer Architecture for HPC, Fortran, Dynamic Allocation in C/C++, Unit Test, Docker, Profiler, HPC Libraries.
• DD2356 - Methods in High-Performance Computing - P4. Topics: Performance Modeling, OpenMP and Shared Memory Programming, MPI, and Distributed Computing.
• DD2360 - Applied GPU Programming - P2. Topics: GPU Architecture, Basic and Advanced CUDA, OpenCL.
• DD2375 - Project Course in High-Performance Computing - upcoming

If you have any questions about the HPC sub-track and the HPC courses, do not hesitate to contact me.

Maxwell's equations are my favorite equations and I teach how to solve them numerically in DD2370 - Computational Methods for Electromagnetics.

I am also an examiner for the MS theses in Computer Science and an advisor for the BS theses.

From 2012 to 2019, I was the co-chair and main organizer of the PDC Summer School at KTH.

Projects

Since 2012, I participate in the major European effort to deliver an exascale system and associated software in Europe. I started with being the programming model work-package leader for the CRESTA project and the scientific coordinator and project manager of the EPiGRAM project. I was also KTH PI for the Intertwine and SAGE projects.

Currently, I am excited to lead as coordinator the H2020 FET-HPC EPiGRAM-HS project investigating parallel programming models for heterogeneous systems (GPUs, FPGA, non-volatile memories, and HBM).

I am KTH PI for

• Sage2: project for parallel I/O and heterogeneous storage systems 
• VESTEC: a project for the visual exploration and toolkit for extreme Computing 

I am co-PI of the SESSI MCP in the SeRC. 

Starting in April 2021, I will be KTH PI for three new EuroHPC projects: DEEP-SEA, IO-SEA and ADMIRE.

I am also chairing the organization of the prestigious 2021 HPDC conference together with Prof. Erwin Laure.

Working with us

Inclusivity is our commitment: we speak about it, we are open-minded and listener, and eager to know you. 

• If you are passionate about HPC and looking for post-graduate opportunities, our HPC group is looking for 2 Ph.D. students in High-Performance Computing. Apply at this link. Don't forget that the deadline April 2nd, 2021!
• If you are a KTH MS student and passionate about HPC, we often have 50% research amanuenses positions. Please contact me for more information about the possibility.