- Use icc
- Use intel fftw
- Simple muti-threading
- vectorize
- check！
- icpc ？？？
- why 3DCTF nx * ny --> (nx+2-nx%2) * ny
- fftw omp critical
- change g++ version
- vectorize for Rebu when g++
- optimize weight、pre bufc and malloc cost
- optimize write
- omp in for z
- rewrite CTF vec 👋
- optimize write2DIm
- New data
see more information in info.md