CERN-HSF logo

CERN-HSF

CERN-HSF (High-Energy Physics Software Foundation) is the umbrella organization for high-energy physics-related projects in GSoC. The HEP Software Foundation (http://hepsoftwarefoundation.org/) facilitates the coordination of common international efforts in high-energy physics software and computing. CERN (European Organization for Nuclear Research, https://home.cern) has participated in GSoC since 2011 as the CERN-SFT group, which provides common software for CERN's experiments. In 2017, the program expanded to include many software projects from the whole field of high-energy physics. The vast majority of our GSoC projects do not require any physics knowledge. The experiments at CERN, such as the Large Hadron Collider, the world’s largest and most powerful particle accelerator (http://home.cern/topics/large-hadron-collider) try to answer fundamental questions about the Universe. For example, what is the nature of mass? What are the elementary building blocks of the Universe? What was the early Universe like? What is the nature of dark matter and dark energy? Why is there an asymmetry between matter and antimatter? In 2012, LHC experiments announced the discovery of a new particle, the Higgs Boson, that helps explain how particles obtain mass. Also, CERN is the birthplace of the World Wide Web. Today, particle physicists are working on analyzing the data from the experiments to study the properties of the newly discovered particle and to search for new physics, such as dark matter or extra dimensions. This requires a lot of sophisticated software. The open-source high-energy physics projects to which students can contribute during GSoC span many high-energy physics software projects: data analysis, detector and accelerator simulation, event reconstruction, data management and many others. We look forward to your contributions!

GSoC Participation History

Technologies

Topics

Past Projects

ML Inference on heterogeneous architectures using SOFIE

SOFIE generates standalone C++ inference code from trained neural networks, letting HEP frameworks run ML models without heavy external dependencies....

Publication of OmniFold Weights A Standardized Format for ML Based Unfolding Results

OmniFold produces per-event weights that cannot be reused without the original analysis environment. The original OmniFold paper deposited input...

Enabling Differentiable Rendering via AD of Parallel C++ STL Primitives in Clad

This project proposes to enhance Clad, a Clang-based automatic differentiation (AD) tool, by leveraging its liveness analysis to automatically...

Fine grained storage for the DUNE experiment

DUNE’s fine-grained Phlex/FORM I/O model stores data products across many independent RNTuple containers, which can increase read-side overhead...

BOA Constrictor: Benchmarking Neural Backbones for HEP Data Compression

The LHC at CERN generates 30 petabytes of physics data every year. Current compression tools like LZMA treat this data like any other file and miss...

EKO Oxidation: Maturing the Rust Core of EKO

EKO (Evolution Kernel Operators) is a foundational Python library used in high-precision particle physics that currently faces performance...

HTCondor Live Tools with Solution Memory

This project enhances the Archi framework by enabling real-time interaction with HTCondor job scheduling infrastructure and introducing a persistent...

Apache Arrow Interface for PODIO

This project aims to implement a high-performance Apache Arrow backend for PODIO. Currently, PODIO data is primarily persistent in ROOT and SIO...

Characterizing I/O Performance for ML Data Loaders at Scale using Darshan

Modern machine learning workflows in High Energy Physics increasingly rely on large-scale datasets, where the data-loading stage can become a primary...

Automated Software Performance Monitoring for the ATLAS experiment

At present, software regression testing at ATLAS is predominantly performed manually via inspection of metrics on the ATLAS project management board....

Clad as a First-Class Gradient Engine in LibTorch

This proposal targets the HSF 2026 idea Clad as a first class gradient engine in libtorch, a 350-hour project intended to make compiler-generated...

Parallel Processing Improvements in Julia Jet Reconstruction

This project aims to improve the parallel scaling of JetReconstruction.jl, a Julia package for jet clustering in high-energy physics. While the...

S3 Backend for RNTuple

RNTuple is ROOT's next-generation columnar data format for High Energy Physics. Its storage layer is decoupled from its format layer, enabling new...

Implement CppInterOp API exposing memory ownership and thread safety

CppInterOp is a Clang-based library that provides an API to examine C/C++ code, enabling language bindings such as cppyy to interact with C++ code at...

Debugging CVMFS Processes in D-State

This project proposes to build diagnostic tooling for extracting debugging information from CVMFS processes stuck in uninterruptible sleep (D-state),...

CARTopiaX: Extending a Next-Generation Platform for Computational Cancer Biology

This project proposes a comprehensive strategy to extend the CARTopiaX platform within BioDynaMo to replicate a wider range of in vitro observations...

AI-Assisted PostgreSQL Optimization Using EXPLAIN Plans

Nopayloaddb is used to serve time-sensitive conditions data for high-energy physics (HEP) experiments, so performance is critical: slow IOV lookups...

Debuggable Installations for Spack Packages

Spack is a dominant HPC package manager that automates scientific software installation, widely used in the HEP community. Spack builds software in...

Data Representation Optimization for ML-based calorimeter simulation

This project aims to develop a pipeline for producing calorimeter shower datasets in point-cloud form for a given detector and various types of...

Integration of Combine with FCCAnalyses

The Future Circular Collider (FCC) is a flagship proposal for a next-generation post-LHC particle accelerator to be built at CERN. As the FCC project...

Enabling Sustainable ARM Support for the Belle II Software

ARM-based processors are becoming increasingly common in HPC environments, as in the AWS Graviton, Fujitsu A64FX (one of the world's fastest...

Generative-AI Assisted Testing of Complex Stacks of Spack Packages

The Problem: Spack’s combinatorial flexibility makes exhaustive CI testing computationally impossible. Current CI focuses almost exclusively on...

Constellation Web Interface for Controlling and Monitoring Experimental Setups

During late-night beam tests, physicists must leave detector halls to find workstations with Qt installed just to check satellite status. This...

Consolidate and advance the GPU infrastructure in Clad

Automatic Differentiation(AD) is a set of techniques to evaluate the derivative of functions specified by a computer program. Clad is a Clang-based...

Transformer based Reconstruction for LHCb PicoCal Calorimeter

Current LHCb calorimeter reconstruction uses rule-based clustering without exploiting timing or learned representations. The PicoCal upgrade will...

AI-Accelerated Signal Reconstruction for the ATLAS Tile Calorimeter at the HL-LHC

The ATLAS Tile Calorimeter (TileCal) currently reconstructs energy using a linear signal-processing method. For the Phase-II HL-LHC upgrade, readout...

Frequently Asked Questions

CERN-HSF | GSoC Org Profile & Stats - Learn about CERN-HSF's involvement in Google Summer of Code (GSoC), their technologies, detailed reports.

|Currently Active|

Contributor Readiness

Participation

Projects

Top Programming Languages

python dominates with primary adoption

Project Difficulty Distribution

Beginner
0
Intermediate
0
Advanced
0

No difficulty data available