Hi All,
I merely wanted to make you aware of this high-performance C++ library for tensor transpositions:
https://github.com/springer13/hptt
I have compared its performance to Eigen: HPTT exhibits significant speedups across a wide range of architectures over Eigen's current implementation.
Please let me know if you are interested in integrating HPTT; I would be willing to assist.
It looks like that HPTT could be an optional path within the Tensor module.
Thanks,
Paul