Summary of GSoC 2021 Projects and Supervisors

Full List of Proposals

Add initial integration of Clad with Enzyme
Add numerical differentiation support in clad
Add support for functor objects in clad
Add support for in-browser interactive averaging of physics results
Add support to resolve symbols from a static library to the Cling C++ interpreter
Allow redefinition of CUDA functions in Cling
Automatic conversion of data stored in TTree form to RNTuple
CernVM-FS preload capability
Deep autoencoders for ATLAS data compression
Developing C++ modules support in CMSSW and Boost
Implementation of physical shape function
Implementing an application for visualizing the LHCb DAQ network
Improve Cling’s Development Lifecycle
Improve the job submission and handling in the Ganga User interface
Improving Cling Reflection for Scripting Languages
MCnet/LHAPDF - Accuracy and parallel computation in parton density calculation
MCnet/Rivet - Modern plotting machinery for the LHC’s MC event analysis tool
MCnet/YODA - Add parallel weight streams to a statistical analysis toolkit
New protocols for exascale data management with Rucio
PODIO serialization back-end for ROOT RNTuple
PRMON - Develop Logging and Unit Test Infrastructure For PRMON
Partitioning GPUs for graphical and computing applications under Linux KVM
Phoenix, interactive data visualization - Development of an experiment independent javascript event display framework and data format
Portability for the Patatrack Pixel Track Reconstruction with Alpaka
ROOT Storage of Deep Learning models in TMVA
RooFit Developmnt - Intuitive Python bindings for RooFit
RooUnfold - Efficient deconvolution using state of the art algorithms
Rucio and CS3API to enable data management for the ScienceMesh cloud
Runtime plugin ecosystem support for OCIS
Scientific Notebook support in ownCloud Infinite Scale
Single-precision floating-point support for GPU acceleration in VecGeom
TMVA Deep Learning Developments - 3D Convolutions for GPU
TMVA Deep Learning Developments - Inference Code Generation for Batch Normalization
TMVA Deep Learning Developments - Inference Code Generation for Recurrent Neural Networks
Upgrading the Ganga graphical user interface
Utilize second order derivatives from Clad in ROOT