Improvement ideas

- [ ] Cleanup incomplete data (like Fedora completed image) if a step is not completed
- [ ] wget: delete/resume
- [ ] Add system info for each benchmark run
- [ ] Better separate variables
- [ ] Rearrange functions to follow the run order
- [ ] Add `--force` mode, in order to force running a step
- [ ] When checking which package(s) to compile, inspect individually
- [ ] Store logs in separate dirs
- [ ] `plot_diagram`: the scaling algorithm is not suited in some cases
- When the right highest Y goes above the left highest Y, diagrams can't be compared
- Idea: normalize all diagrams with a (perfect scaling) line that goes from 1 thread to nproc threads

The issue lists multiple improvement ideas for a shell script that runs the PARSEC benchmark suite in a RISC-V emulation environment. Start by examining the main script to understand the step order, variable usage, and logging structure. The 'plot_diagram' scaling issue likely relates to a separate plotting script or function. Implementing any single item requires understanding the existing codebase and the PARSEC benchmark workflow.

Vom Indexierungsmodell aus dem Issue-Text verfasst.