Hi, i read what you're saying about core x clock, but at this point i feel a little confused.
I work major time with corona renderer, so im looking for new upgrade for workstation and based on i read here u re saying that is better to have a 2 x Xeon 6 core X5675 3,1 GHz vs 2 x Xeon 8 core E5-2650 with 2 GHz ? Is wise to do the calculation based in nºcores vs clock?
X5675 ->12 c x 3,1 Ghz =37,2 Ghz
E5-2650 -> 16 x 2GHz = 32 Ghz
The best choice is the X5675 for this case (corona render only)?
Thanks
No one is saying that :- ). The mathematics is absolutely sound but more complex:
[All core turbo frequency] x [ amount of cores ] x [ generation multiplier ]
The mathematics work right out of the box only when comparing exact generation ( Haswell-E vs Haswell- E ), but not so much when different generations, like in this case first-generation SandyBridge-E vs old Westmere-EP.
Also, it's sometimes hard to find out the all-core turbo frequency, because Intel only writes single-core turbo in their description, but wikipedia lists all the multipliers.
Anyway, it's supposed to be just general estimate, I personally always use Cinebench R15 scores.
I don't really have time to find the numbers for your two particular CPUs, but although the odds are against the 2650v1 because of low-clock (Max single turbo is 2.8 and I think the all-core might be like 2.3), SandyBridge was big upgrade from Westemere, so it will probably hold up both in single and multi-core performance just fine. Might even come up better.