GitLab

Adjusted mapping selection logic in tensor scale function.  God I hate diagonal iterators and scalars.

Implemented permuted slice. Working for multiworld reads/write. -1s not tested yet, and same world fast-redistribution not
hooked up, though implemented.

Also, simplified interface.

Addec a couple of unit tests and turned on diagonal rescaling. SCF may be incorrect currently though

Implemented a kernel that does a distributed reshuffle on the virtual block level when possible.

Deprecated C interface for cyclopstf.hpp. Added support for element-wise custom operations. Got rid of buffer and buffer_len parameters that were never used.  Added a test for weight
which uses a custom divide operaiton.

folded dimensions were set to edge length 1. Fixed by changing symmmetry of
folded dimensions symmetry to NS in the lower level kernel.

 Added compare_tensors function to print two tensors side-by-side and changed print_tensor to only print non-zero elements.

Wrote code that explicitly zeros out the padding. Made CTF_World be object reference rather than pointer everywhere.
Having issues with FFT tests due to diagonal rescaling in SY and possibly other problems.