Better performing doing what ? Reference or non-reference ?
Gaming ?
Above we see the reference Titan X outperforming the 980 Ti by about 4.6% (91/87)
But the Gigabyte G1 980 Ti overclocks a whopping 30.8 % (134.8 / 102.6)
http://www.techpowerup.com/reviews/Gigabyte/GTX_980_Ti_G1_Gaming/33.html
Techpowerup's review of the Titan X managed just a 11.8% improvement (191.7 / 171.5)
Given the extreme overclockability of the 980 Ti's, and the relative pricing, I wouldn't use anything else for Gaming or CAD workstation boxes. Jump into CAD Rendering and the Titan X would be the better card.
The RAM I see as immaterial.... tho it's possible to use more than a particular card offers, I have never see a performance impact by dropping below that level.
http://alienbabeltech.com/main/gtx-770-4gb-vs-2gb-tested/3/
As shown in the test above, games showed no overall improvement using 2 Gb vs 4 GB 770s. Even a game that used 2.75 GB of RAM on a 4 GB for example, will not experience any slowdowns, dropped frames or other negative impacts if you swap in a 2 GB card.
The same story has come out each time this test has been performed since.... poorly written console ports can however be the exception to this rule.
https://www.pugetsystems.com/labs/articles/Video-Card-Performance-2GB-vs-4GB-Memory-154/
http://www.guru3d.com/articles_pages/gigabyte_geforce_gtx_960_g1_gaming_4gb_review,12.html
At 4K, the answer I think is immaterial. As the Titan / 980 Ti com up well short of 60 fps on 4k in current AAA Games, I think we have to wait a generation or two before this matters.