[zero_custom]

Hi all,
I've been looking into saving and building a computer for simulations and possibly rendering. I'm at the fork road as deciding a CPU very much decides the "future proving" of your system. I'm interested to hear people's thoughts and experiences with them or troubles / stability of overclocking or anything that I may not have experience with.
I've a few choices of CPUs but they all have their pros and cons using sites for comparison.

in no order
xeon E5-2680V3 ( 12 cores and has a clockspeed of 2.5GHz)
xeon E5-2667V3 ( 8 cores and has a higher 3.2GHz, cost the most)
i7 5960x (8 cores but it has a higher clockspeed of 3.0GHz)
i7 6700k ( 4 cores of highest clockspeed of 4.2Ghz and cost the least)

http://ark.intel.com/compare/83361,8...08,88195,82930

If there's a "better" CPU or one of the 4 I shouldn't get, or even a cpu I should wait for in the coming months. I'm all ears.

[flotus1]

The 4 CPUs you selected have an entirely different scope of application.
Choosing the "right" one(s) depends your application (mainly the problem size) and your budget.

[urosgrivc]

Lets say I want to mesh 4 parts for simulation in cfx,
won't the i7 skylake win? or 5820k OC ? I meen we are talking about 5 time cheaper cpus here that on paper will mesh quicker if we dont have more parts than 4 or 6.

Am I wrong?

i7 skylake Should also open ansys quicker and load results quicker. As for single threaded operations it is much faster than any E5.
i meen do I need two computers? one thats fast (i7 or E3) and one that has lots and lots of cores (dual E5)?
What I m trying to say is that I see my computer working a lot in single thread, always when i am waiting one thread is at maximum.

Zero_custom - for rendering, I would go for more cores with Hyper-threading as rendering is usualy a multythreded operation. i7 6700k doesent stand a chance here.
and for simulations well i have described a problem abbove.

[Edeluc]

Originally Posted by urosgrivc

Lets say I want to mesh 4 parts for simulation in cfx,
won't the i7 skylake win? or 5820k OC ? I meen we are talking about 5 time cheaper cpus here that on paper will mesh quicker if we dont have more parts than 4 or 6.

Am I wrong?

i7 skylake Should also open ansys quicker and load results quicker. As for single threaded operations it is much faster than any E5.
i meen do I need two computers? one thats fast (i7 or E3) and one that has lots and lots of cores (dual E5)?
What I m trying to say is that I see my computer working a lot in single thread, always when i am waiting one thread is at maximum.

Zero_custom - for rendering, I would go for more cores with Hyper-threading as rendering is usualy a multythreded operation. i7 6700k doesent stand a chance here.
and for simulations well i have described a problem abbove.

Before choosing the CPU you should think a little bit more about your problem in general. If you plan to do CFD simulations, your workstation layout might be very different to a workstation layout that is designed for e.g.: Abaqus simulations.
Usually you should have a look at more than the CPU component and also consider the amount of RAM and which type of storage (HDD, SSD) you want to use.
For CFD simulations you usually need more RAM but you do not necessarily need SSDs to speed up the simulation, because you do not perform much I/O. Also think of your operating system. Linux can use the RAM to perform very fast I/O unlike windows.
What I am trying to say is that only thinking of one component might cause some bottlenecks in the overall system, so think of the "bigger picture".
As I saw that you mentioned ANSYS, here is a useful guide for designing your workstation for FLUENT:
https://storage.ansys.com/corp/2012/...t/it_guide.pdf
This pdf covers the things above pretty good.

To give a more concrete statement regarding the CPU's - If your budget permits it, I would go for the XEON series, because in comparison to the i7-6700 series they have way more L3-cache and they have QPI links, which might be helpful.
Again, this is my personal experience and might not be the best choice for you.
Browse the web some more maybe even ask the user support of ANSYS what they might recommend - They know their code best and therefore know the optimum hardware.
I hope this helps.