Hi All, With the help and support from Edmon, I am going to leverage the HPC resources at NICS ( https://www.nics.tennessee.edu/ <https://www.nics.tennessee.edu/>) for some Drill tests. The goal is to explore how Drill performs on HPC platform.
1. Platform configuration: details can be found here: https://www.nics.tennessee.edu/computing-resources/beacon/configuration <https://www.nics.tennessee.edu/computing-resources/beacon/configuration> 2. Tests: TPCH and/or TPCDS (subset of queries) datasize: will start with SF1000 (~1TB), may increase to larger dataset (depending on storage space available) test runs: 1 thread (like power run) multiple threads (like Throughput run) 3. Set up: — Drill (master) on Apache Hadoop — Drill (master) on MapR-FS Any comments/suggestions? Thanks, Dechang