ulugbek-abdimurodov/EvaluatingAutomatedMethods4MicrobenchmarkGeneration ? reverse-engineered prompt

Reverse engineered prompt

Build me a clean research repo for a study about comparing two ways of generating microbenchmarks for Apache Ignite, one using an LLM on its own and one using JUnit tests first, then converting them into JMH benchmarks.

I want the repo to help someone reproduce the experiment end to end, from checking out the right Ignite commits, rebuilding the project, generating the tests and benchmarks, running the benchmarks, and comparing the results between the buggy version and the fixed version. Include simple scripts for cloning, switching commits, building, generating benchmarks, and running them.

Also set it up so the important data files, analysis results, diagrams, and thesis document are easy to find. Make the README clear enough that a non expert can follow the workflow and understand what each folder is for. If anything needs current docs or setup details, look them up online and keep the instructions practical.

Are you gonna build this?

make sure you review the code using coderabbit

Try freeSponsored — opens CodeRabbit in a new tab