In this repo, you'll find all the data to recreate the results shown in the SoCC 2020 paper "Specializing the network for scatter-gather workloads"
- cpu_sender: All scripts representing the different CPU baselines: incast tests, static batching to solve incast, and the rate algorithm implementation.
- worker: All workers used for the experiments. They are common for both the FPGA and CPU experiments.
- ns3: Simulator used for the scalability tests.
- plotting: plotting scripts.
- docs: Documentation.
The FPGA sender implementation can be found on its own repository.
- Catalina Alvarez - ETH Zürich
- Zhenhao He - ETH Zürich
- Gustavo Alonso - ETH Zürich
- Ankit Singla - ETH Zürich