In order to find the pixel location from GCP you need to be able to trace a ray from the ground to the camera and see where it intersects your image. That GDAL can't do, as it does not know anything about how the Digital Globe camera works.
On the supercomputer if a GUI is needed I usually start vncserver (if yours has it) and then connect to it with vncviewer (via port ssh forwarding). Or, you can copy your images back from the supercomputer where you have a GUI, do the work there, and copy things
back (but DG images are so huge). Maybe you can just make a little DEM crop, mapproject the DG image onto that crop, and copy the resulting little orthoimage over locally with rsync, then for each pixel you will have a lon-lat.