Thanks for the feedback Sonny,
I've been checking some cinebench scores a few weeks ago.
What I see from http://www.cinebench.info is these screenshots:
Dual Xeon's E5-2620 v3: score 1553 [almost 18% faster]
i7 5960x: score 1318 [only 3% faster!]
i7 5820k: score 1275
So, the dual Xeon setup has a clear advantage (in benchmark).
And, the 5820k is almost touching the 5960x.
The only concern is that cinebench renders a not so complex scene that is relatively fast comparatively to real vray cpu rendering still images that take hours to complete. So if a 5960x is only 3% faster by cinebench scores then the results pretty speak for themselves.
On the other hand, if we look to cpu-monkey.com the scores are far more apart from this previous conclusion:
5960x [on Cinebench R11.5, 64bit (Multi-Core)] scores 13.18 [almost 20% faster than 5820k]
5820k [on Cinebench R11.5, 64bit (Multi-Core)] scores 11.05
E5-2620 V3 [on Cinebench R11.5, 64bit (Multi-Core)] scores 9.1 - which would double to 18.2 ??? [that would be 38% faster than 5960x??]
Or
5960x [on Cinebench R15 (Multi-Core)] scores 1360 [25% faster]
5820k [on Cinebench R15 (Multi-Core)] scores 1085
E5-2620 V3 [on Cinebench R15 (Multi-Core)] scores 894 - which would double to 1788? [meaning 31% faster than 5960x]
I know that these CPU's ain't that far appart regarding performance, but money doesn't grow in the trees these days. Just need the best bang for the buck. I think i'll just settle for the 5960x: they are readily available, are much more future proof, and i'm leaving the option opened for Vray RT down the line with a GTX970 that can be coupled latter IF I'd get a real performance increase using SLI.
Better still, I think I could overclock the 5960x to 4.3 to 4.4Ghz easily using Asus software.
Any sense?
PS: Sorry, but I think my posts are being moderated so there's a quite significant delay between my posts.