both ends i guess, but the back end emission is dumped in the metal case. the beam emerges in the interface zone between the narrow uppermost bar and the thick slab underneath.i have tried to separate both with little success. however, in the tiny specks of debris i noticed that there is what looks like a groove along/under the upper bar, probably a wave guide where actual lasing occurs /see the pic below/ i wonder what a multimode diode would look like in that region.