San Diego Supercomputer Center · UC San Diego
Research
My work spans high-performance computing related topics in structural engineering, plasma physics, genomics, high-energy physics and astrophysics. Below are the projects I'm currently most active in — follow the links for details, code, and publications.
Current Projects
SHAPE Center
Scalable HPC Applications for Plasma and Energy
Co-PI of the SHAPE Center, developing scalable high-performance computing methods and applications for designing metamaterials under high strain.
Learn moreCGYRO
Gyrokinetic turbulence simulation code
Core developer of CGYRO, a high-fidelity gyrokinetic turbulence simulation code used to model plasma microturbulence in fusion devices. CGYRO runs on the world's largest supercomputers and is a key tool in the fusion energy research community.
CGYRO docsscikit-bio
Bioinformatics algorithms for Python
Moderator and contributor to scikit-bio, an open-source Python library providing data structures, algorithms, and educational resources for bioinformatics. The library is widely used in microbiome research and computational biology workflows.
scikit.bioCosmos
APU-accelerated computing at SDSC
Co-PI of Cosmos, SDSC's APU-based computing system. In addition to co-leading the project, I help users port their codes to APU compute, enabling them to take advantage of the system's unique heterogeneous architecture.
Cosmos sitePast Projects
IceCube Neutrino Observatory
High-energy neutrino detection at the South Pole
Contributing to the IceCube collaboration, the world's largest neutrino detector buried in Antarctic ice. My work focused on the computing infrastructure and data processing pipelines that handle the observatory's high-throughput data streams.
IceCube siteOSG
Open Science Grid — distributed high-throughput computing
Contributed to the Open Science Grid, a national cyberinfrastructure that provides distributed high-throughput computing resources to researchers across the US.
osg-htc.orgNRP
National Research Platform
Worked on the National Research Platform, a distributed computing infrastructure connecting research institutions to enable large-scale data-intensive science.
nrp.aiglideinWMS
Workload management system for grid computing
Developer of glideinWMS, a workload management system that enables efficient job submission and resource brokering across heterogeneous grid and cloud computing environments.
glideinWMS docsSelected Publications
Full list on ScholarA full list of publications is available on Google Scholar and ORCID.
A. Marinov and I. Sfiligoi, “Structuring agentic AI for HPC code modernization”
USRSE26 https://doi.org/10.5281/zenodo.22283502
I. Sfiligoi, E. A. Belli and J. Candy, “Minimizing CGYRO HPC Communication Costs in Ensembles with XGYRO by Sharing the Collisional Constant Tensor Structure”
ICPP Workshops '25 https://doi.org/10.1145/3750720.3757303
I. Sfiligoi et. al., "Optimizing UniFrac with OpenACC Yields Greater Than One Thousand Times Speed Increase"
mSystems7:e00028-22. 2022 https://doi.org/10.1128/msystems.00028-22
I. Sfiligoi et. al., "The Pilot Way to Grid Resources Using glideinWMS"
2009 WRI https://doi.org/10.1109/CSIE.2009.950