Run FreeFEM with single precision

In the changelog of version 4.6 it is written that FreeFem can solve problems with single precision or 64 bit integers. How can a FreeFem script be run with single precision?

I would like to use this feature to observe the influence of roundoff errors, without the need of using too fine meshes.

Thank you,

This only works with PETSc/SLEPc. You need to have an installation of PETSc/SLEPc with the additional configuration parameter --with-precision=single. Let me know if you need assistance with that.

Thanks for the quick reply. I will think about having a separate installation with single precision.