The first call for short term collaboration was launched in 2018, inspiring the 4 following projects:

• HPC for Neutron Transport in a Fusion Reactors Core: From Comisión Nacional de Energía Atómica, CNEA, Argentina; to Barcelona Supercomputing Center (BSC)

• Scaling Large Machine Learning Models for HPC: From Barcelona Supercomputing Center (BSC); to Tokyo Institute of Technology
  Parallelism in deep learning training is inevitable. Mostly because of the need to accelerate training and recently due to larger models and bigger datasets. Though data parallelism is the most common parallelisation method, numerous strategies have been suggested, each with a promise to accelerate the training process. However, there is no clear way to determine which approach to use for a given model or dataset. By analysing memory, communication and performance, we have developed an oracle that suggests the best parallelisation strategy given a set of parameters such as model, dataset and system specifications. We have also performed a cocktail of tests with standard datasets and with newer emerging scientific datasets such as Cosmoflow on the ABCI supercomputer.

• Math Libraries Migration and Optimization on ARM-based processors: From Barcelona Supercomputing Center (BSC); to Fujitsu

• Processing in Memory for Data Intensive Applications: From University of Malaga; to ETH Zürich
  Thanks to the cross-pollination of our research, I obtained more than 3.6x better performance compared to the general-purpose cores approach,  using a PIM approach rather than a typical host configuration. Using the custom accelerator based on RTL logic, the speedup reached 18x. Those results, including the energy consumption estimation that we obtained, is outlined in a conferencepaper, which will be submitted to HPCA.

• Themis high performance simulation framework: From University Grenoble Alpes/Inria; to Imperial College

The second call for short term collaboration projects was launched Spring 2019 and resulted in the granting of these 6 applications:

• Energy efficient execution of tensor application on heterogeneous platforms through cross-stack optimisations: From Chalmers; to TU Dresden

• Multiplex-heterogeneous network embedding for drug repositioning: From Institut de Mathématiques de Marseille; to BSC Lopez Léo

• Cloud-based, manycore platforms for realistic brain simulations: From Erasmus Medical Center; to NTUA

• JIT code patching on low-power processors: From University of Manchester; to FORTH

• Exploring coevolutionary methods to determine interacting partners in protein systems: From Universidade de Brasilia; to BSC

• Resilient Deep Learning Training on Massively Parallel Computers: From Costa Rica Institute of Technology; to BSC

A Eurolab4HPC special Short Term collaboration grant launched in Fall 2019, resulted in these funded projects:

• Toward a New Generation of Sparse Solvers: From University of Illinois at Urbana-Champaign; to Imperial College

• Searching for the optimisation strategies of HPC software: From University of Münster; to University of Edinburgh

• Eurolab4HPC Short Term Collaboration Grant Proposal: From NTUA; to Inria

• Innovative HPC cooling: From Politecnico di Milano; to EPFL

• Evaluation of genome sequencing workloads using SVE: From BSC; to ARM

• A Fast and Composable Simulation Infrastructure: From NTNU; to EPFL

• On Resilience of Stochastic Computing under Aggressive Undervolting: From University of Teheran to BSC