I personally almost never compress my sub. If I do compress, I do so for more of a tonal sound vs volume intentions, which I will mix with the dry signal of the sub.
I compress after EQ - just logically for me, it's less work for the compressor if you've already chopped out the muddy and offending frequencies.
I don't know if other people do it this way, I always suggest learning by trial and error. Try it both ways - mix it both ways to the extreme, and in more conservative ways and play around with your results.