CoercionShaman
Your comment about having micro stutter would indicate that the Z is a bad choice.  You obviously notice it.  Yes, some of it may be memory issues, but it is often SLI related.
 
Also, the performance of the X is shown to be better than the Z and on a single chip in benchmarks and on the NVIDIA website for 3DMark Fire Strike.
 
http://www.geforce.com/ha...tx-titan-x/performance

ant4423
...one powerful GPU is the way to go so I def say 1 Titan X.

1. Raw Power: Theoretically, if sli scaling works as expected for a particular game, the TITAN Z should beat the TITAN X by a good margin. Why? Because the Z has a total of 5760 keplar cuda cores (shader processors) while the X has only 3072 cuda cores. But the keplar cores are running at around 900Mhz (because two GK110Bs in a small space give out too much heat) while the maxwell cores are running at 1300Mhz at full load. So, it's not a one-keplar-core=one-maxwell-core.
2. Real world Comparisons: There are currently only two reviews sites that have directly compared the Z and the X and they show clearly that the Z beats the X. &nbsp www.guru3d.com/articles_pages/nvidia_geforce_gtx_titan_x_review,15.html     www.kitguru.net/components/graphic-cards/zardon/nvidia-geforce-gtx-titan-x-12-gb-sli/13/       (Please remember, this is only when SLi works to a good degree. A dual core card will be inconsistent in beating a single core across many titles.) Yes, the Z beats the X but not by much but it still beats it, strictly speaking. Also, in all of the comparisons in the two links, the Z does beat the X always. However, Nvidia's own webpage shows the X beating the Z on 3dmark 
3. Micro-stutter: When I was on the Z (and titan og SLi before that), I had a healthy dose of microstuttering in many many games. I was quite irritated with it. I blamed everything but SLi and it was only when I switched to the X that I realize how much important frametime consistency is. But when SLi did work properly in some games, it was perfect, and I cannot say it was all bad. Anyone who says micro-stuttering was cleared out 10 years ago doesn't know what micro-stuttering is; G-sync won't solve it and sometimes exacerbates the effect. There was a huge controversy a couple of years ago regarding AMD CF setups not at all using the second card because of largely runt or dropped frames. Following that frame time variance and consistency took the foreground over raw performance. SLi and CF still do have the consistency vs performance tradeoff.
4. future proofing wrt W10, DX12: Assuming DX12 gets quickly adopted, future titles could be using one of these three DX12 multi-adaptor features: Multi-adaptor implicit (basically driver level control meaning SLi/CF with profiles needed for new releases), Multi-adaptor explicit linked (requires two same cards, stacked memory and each gpu handles different work), Multi-adaptor explicit unlinked (any gpu, work load between gpus decided by developer).  Source:  www.pcper.com/reviews/Graphics-Cards/BUILD-2015-Final-DirectX-12-Reveal   Optimisitically, if the second option became prevalent, the Z would become a powerhouse with 12GB stacked memory and 5072 cores and could see it topple other cards in performance. But that is a BIG IF. Moreover, while most keplar and fermi cards will support DX 12, ALLL of the features of dx12 won't be supported and only the X and future cards can do that.

GODLIKE99KINGZ
titan x holds HDMI 2.0 or HDMI 1.4? My titan Z runs with v.1.4, I wonder if X been upgraded to v2.0