One thing to keep in mind is that the Tri-X you buy today is not the same as the Tri-X your father (or grandfather) came to know and love. When Kodak moved production of this film to a new coating plant a couple or three years ago, significant changes were made. The film still has the tonal range that made it famous, but the new version has less grain and from what I can tell the emulsion is more resistant to damage when wet. IMO, it's a good thing though I'm sure others' opinions will differ. So if you are looking for the "Tri-X look" of ten, twenty, or thirty years ago; you might want to look at a different film. FortePan 400 comes to mind as one of the available film stocks with much more pronounced grain. The film is also available as Arista.edu "Made in Hungary" from Freestyle Photo. I've used this film in 120 and 4x5 and it shows a lot more apparent grain than and has a completely different look to it than Tri-X. Ilford's HP5+ is another good choice.
I've processed a ton of Tri-X in XTOL diluted anywhere from 1+1 to 1+3, D-76 straight and diluted 1+1, and a much smaller boat load in Rodinal at 1+50. I'm not a grain fanatic, so I've pretty much crossed Rodinal off my list of developers suitable for Tri-X. Properly done, you can get a very nice tonal range with this combination along with very sharp and well defined grain. I haven't come across anything else that produces quite the same look. Occassionally, I'll use this combination for that particular effect though not often. XTOL is at the other end of the spectrum; very sharp, very smooth tonal range, a bit of a speed boost, and very fine apparent grain. Once diluted, the results are pretty much the same regardless of the ratio. Development times are longer at higher dilution ratios, but that's about it. D-76, if you dilute it 1+1 is somewhere in the middle. Apparent graininess is less if you use it full strength, but what is there tends to get a litle muddy looking.
There is quite a bit of published information about using any of the developers I've mentioned with Tri-X. If lots of experimentation is not to your liking, the numbers published on "The Massive Dev Chart" will get you reasonably good negatives right from the start. Getting your personal preferences down exactly should take no more than 3 or 4 tries from there.