This is a very preliminary version of a code generator for tensor transposes. The intended application is the coupled-cluster codes in NWChem, but any related program which does quantum many-body computations may benefit from it.
It will be some time before I attempt to make this code useful for anyone else. That means there is no technical support and no documentation, nor do I even expect this code to work for you.
Jeff Hammond, July 2008
This repository also includes Fettuccine (Fast thrEaded Tensor Transposes Using C and Cuda In Nvidia devicEs), which is a combined effort by Tom Henretty (OSU) and myself.
Jeff Hammond, August 2008
While I'm far from finished with the transpose library, I'm starting to work on true native-dimension tensor contractions instead of transpose-transpose-dgemm-transpose. Consistent with the theme, this library is going to be called Macaroni (Multidimensional Array Contraction...)
Jeff Hammond, September 2008
Use Patent Claims
Include Install Instructions