generate codes with retrieval or without retrieval

python main.py --dataset polybench --method generate_multi --use_rag True 

compare results

python compare_result.py --dataset polybench 

count original checksum and execution time of datasets

python count_oritime_and_checksum.py --dataset polybench 

count execution time of pluto_optimized versions of datasets

python count_plutotime.py  --dataset polybench 

Don't forget to fit in your api-key in llm_response.py!