The Accelerated Big Data Systems (ABS) group is a research group formerly active at the Delft University of Technology since 1999.

The ABS group performs research into high-performance computing infrastructure to enable the efficient processing of big data problems. A number of research topics are carried out in the group, mainly focused on enabling next generation high performance computer hardware to perform scalable big data analytics, such as addressing the limitations in parallelism on a scalable computational infrastructure. In addition, we work on enabling efficient utilization of computer hardware and enabling easy integration of GPUs and FPGAs into big data frameworks.

Research in the field of accelerated big data analytics is mainly industrially driven to address real-life challenges in the field. A couple of application domains our group focuses on is the field of genomics as well as analytics. These are becoming strongly data-driven due to the abundance of extremely large datasets and the high complexity of the analysis algorithms. These efforts have resulted in multiple high-tech spinoffs such as Bluebee a genomic analysis company and Tertide a high performance big data analytics company.

This GitHub organization contains various projects formerly or currently developed within the ABS group. Some projects also live in people's individual GitHub spaces.

Projects are either an accelerator/acceleration implementation for some computation task on a certain platform, or a tool. This tooling can be a code-generation tool or debugging tool for designing hardware, or a runtime tool to interface with an accelerator.

For documentation of projects related to Tydi and debugging in Chisel (Tydi-Chisel, Tywaves, ChiselTrace) and Clash (Tydi-Clash, Shockwaves), check out https://abs-tudelft.github.io/docs/.

Feel free to get in contact with the authors of any of the repositories by opening an issue on the repository or shooting them a message.