I have not yet found the article I had read, but plenty to find there. The article I read irradiated the entire thin slab top and bottom, but put a high reflectance mirror at a slight angle on one end, and an output coupler mirror on the other. The beam inside the crystal ran through it to the HR mirror and then back through it again at an angle through the crystal before exiting through the OC or output coupler mirror. Whether the mirrors were discrete output mirrors or deposited upon the ND:YAG crystal itself I do not remember, but recall the crystal had a slanted cut on one or more sides. Saw the article with diagrams several years ago.
It has been too long to remember the wavelength of the diode arrays, but for 1064 nm output surely had to be IR and I believe they used 50 or 100 watt square arrays on each side to pump the crystal with. I may have this article saved somewhere, if you really want it, I will spend some time searching through my files.