Does 7xxx series support atomic float operations so AMD could expose GL_NV_shader_atomic_float extension?
Is AMD working on exposing it?..
Unfortunately, current AMD hardware cannot perform atomic operations on floating point data and so we cannot expose GL_NV_shader_atomic_float. Operations that can be expressed using integers (such as logical operations and compare-exchange) can be emulated using integer format images. However, operations such as floating-point addition and subtraction cannot be made available.