Integral direct OpenMolcas
One of the most significant things holding back OpenMolcas towards massive parallelization and calculations on large systems 1000-10,000 atoms is the I/O. In this respect I would like to ask if there is any interest in joining me for a hackaton at which we start by addressing this problem. The first thing I would like to achieve is to make all the one-electron integrals direct and computed on the fly as they are needed. Additionally, the presence of other files might also be questioned.
Once this is accomplished I would like to see the a collective move to make the two-electron integrals direct too.
Anyone interested in joining?