scripts/validate_schema.py stays at the top level (most-used, referenced everywhere).
All other scripts are organized into functional subdirectories.
| Directory | Contents | Usage |
|---|---|---|
generators/ |
Spec generation scripts for each benchmark suite | python3 scripts/generators/generate_rodinia_specs.py |
survey/ |
Codebase surveying and kernel inspection scripts | python3 scripts/survey/survey_rodinia.py |
analysis/ |
Results analysis and report generation | python3 scripts/analysis/generate_report.py |
baselines/ |
Baseline population scripts | python3 scripts/baselines/populate_baselines.py |
augmentation/ |
Augment→build→run→verify pipeline | python3 scripts/augmentation/augment_verify.py specs/rodinia-bfs-cuda.json |
batch/ |
Shell batch runners for multi-spec testing | bash scripts/batch/run_rodinia_batch.sh |
archive/ |
One-time fix scripts (kept for reference, not for re-use) | — |
validate_schema.py— JSON schema + cross-cutting validator (path unchanged from docs)augmentation/augment_verify.py— single-spec augment pipelineaugmentation/run_augment_batch.py— batch runner: spec list × augment levels → JSON + MDaugmentation/combine_aug_results.py— merge multi-stream batch resultsgenerators/generate_rodinia_specs.py— Rodinia spec generatorbatch/run_phase_cuda.sh,run_phase_omp.sh,run_phase_opencl.sh— full augmentation test runs