Data transport and correlation

The main objective of this workpackage is to deliver a collection of efficient, and/or high-performance signal processing building blocks exploiting commercially available hardware accelerator platforms (FPGA, GPUs) that partners can use to construct a next-generation correlator (the machinery that combines signals from individual detectors to synthesize a more sensitive instrument that also has greater resolving power) capable of processing for the targeted facilities much more data, be it driven by bandwidth (EVN, e-MERLIN and ALMA) or field-of-view (LOFAR, SKA-VLBI). The project will integrate these common components into a demonstrator; the outcome of the workpackage will be used later by each facility and upcoming instruments (like the ngEHT) to upgrade their correlators, such that they meet the high data requirements delivering the science of the coming decade.

The workpackage includes partnership with Sioux Technologies, to use the same AI technology for their platform for deployment, with mutual benefits because of the exchange of different technologies. ESO, KASI and NAOJ contribute in-kind to this work package because of their interest to use tensor-core GPUs for the next generation correlator for the ALMA telescope. They will work in close collaboration with the partners to specify and develop the tools required.