Rouxinol

Machine learning has greatly improved compiler technology. Its effectiveness relies on accurate and expressive code representation, influencing the model's learning ability and performance. Rouxinol provides developers with an infrastructure to explore and innovate in this field. It is the tool developed for our CC 2025 paper, with reproducible results detailed in CC2025_Artifact.pdf.

Anderson Faustino da Silva, Jeronimo Castrillon, Fernando Magno Quintão Pereira, "A Comparative Study on the Accuracy and the Speed of Static and Dynamic Program Classifiers", Proceedings of the 34th ACM SIGPLAN International Conference on Compiler Construction (CC 2025), Association for Computing Machinery, New York, NY, USA, Mar 2025.

---------------------

NOTE: 

1) Line 219 of the file /.../Rouxinol/artifact/cc2025/data_generator.sh contains an error. Replace

compiler_flag="-mllvm -fla -mllvm -sub -mllvm -bcfe"

with

compiler_flag="-mllvm -fla -mllvm -sub -mllvm -bcf".

2) Line 166 of the file /.../Rouxinol/artifact/cc2025/data_classifier.sh contains an error. Replace

for (( r=0; r<${#rounds[@]}; r++ )); do

with

for (( r=0; r<${rounds}; r++ )); do