After spending a better part of my night looking through Nvidia press releases and the like about the 1080, I was unable to find out what I was looking for. Since the 1080 has double floating point precision like the titan X, but seems to be better than the titan x in many ways, I was thinking of upgrading. What would you say the likeliness of Nvidia and graphic processing programs supporting the product line for gpu accelerated tasks? While I know geforce cards are not workstation or server quality, I am looking for something for my home rig for when I work at home. Other than memory, the 1080 cards seem to mostly be better performers than Tesla cards. The 1080 cards also have better bandwidth, lower power consumption, and can be replaced multiple times before equaling one Tesla card.
An example of use will be running Navier Stokes solver in ANSYS Fluent. This software will allow the setup of GPU acceleration with any cuda core card, but will not actually utilize them unless they have been supported by ANSYS / Nvidia (sometimes "tested" cards work I have been told).