Zhisong Qu

What @zhucaoxiang suggested could be one of the problem. The source of randomness may also come from the fact that I randomized all the field variables(initial guess) to some small...

@jloizu That was in the latest commit of branch helicity-fix. I have add the following lines in preset.h ```fortran call random_seed() call random_number(Ate(vvol,ideriv,ii)%s) call random_number(Aze(vvol,ideriv,ii)%s) if (.not. YESstellsym) then call...

Following what @SRHudson suggested, I have made the random initialization an option rather than compulsory. Now if we set Linitgues=3, the initial field will be randomized. The maximum of the...

We could merge this branch. However, the hessian for toroidal geometry needs further development.

Done. I have also screened warning messages for the Newton's method. I suggest we continue this thread for future work of Igeometry=3.

I am happy to announce I have a version of SQP that at least "seems to work". I have removed all the Lagrangian multipliers in dMA, and moved them to...

Sorry my mistake. So we have _df00ab_ __denrgy__ compute the derivative of energy in each volume for field Newton method _tr00ab_ __iniota__ compute the rotational transform on the interfaces

@abaillod Thank you for your impressive progress. After some search, it seems to me that `!$OMP CRITICAL` makes the threads queuing to use the FFTW library. The website you quoted...

@abaillod I found another possibility that slowed down the calculation, that is, the LAPACK subroutines that inverts the matrices. If you are using Intel MKL libraries, you should be able...