I am debugging cpp backend code generated for gcc. I took help from page Inductor CPU backend debugging and profiling — PyTorch Tutorials 2.6.0+cu124 documentation and i was able to generate output.py
I need to peek into compilation process so i have TORCH_LOG=all set which provided me with compilation command. However I face issues for header files , also mentioned in documents mentioned above.
#include "/tmp/torchinductor_root/lr/clrlgu27q4ggd472umdzwsu6qcpqxcuusjxqvx2hwitjbujiiz7z.h"
How do get access to this files in compilation.
It would be really helpful if you can tell more about compilation flow using intermediate files. e.g. python3 output.py // to generate kernel binary files directly from g++
Thanks
Mahesh