Best processing speed for fluent -2600k or 3570k?
 

Hey Guys!
I'm an undergrad and I am planning to work on projects involving the use of fluent quite frequently. So I'd like to know which of the above listed processors would be better in terms of iteration speeds? Does anyone have benchmarks for these? Also, the price difference in the market where I live between these two is around $35. So I don't mind going for the 8 threaded i7. But will the lower TDP of the IB i5 really help in reducing power consumption and temperatures? And is it worth giving up hyper threading in the 2600k? And I may use this one for gaming too although w/o a video card? So will HD 4000 w/o HT be a better choice?

I'll really appreciate some help. Thank you :):)

The newer i5 chip has a slightly smaller cache, which may or may not be offset by the slight architecture efficiencies gained from Sandy Bridge to Ivy Bridge.

The on-chip GPU for the i5 is significantly better, but it's still nearly worthless. A $35 nvidia or AMD pci-e card would destroy it.

Hyperthreading doesn't matter. Not for games because most of them don't use 4+ threads effectively, and not for CFD because the bottleneck is memory bandwidth and cache performance.

The power savings of the i5 are negligible, and it actually appears to be a little worse as far as thermal management. The chip itself is smaller so there is less surface area to conduct heat away to the cooler. Intel also skimped on the heat spreader for Ivy Bridge. It used to be soldered on with Sandy Bridge, now they just use an epoxy that isn't as conductive.

Other than that, the overclocking is about the same, they can use the same speed memory, and they can use the same chipsets.

Verdict: It don't matter. Buy the cheaper one.

Hey Kyle,
Thanks a lot for the input. :) :)
I'll mostly go for a 3570k then and buy a cheap GPU for casual gaming. Also, should I keep anything in mind while buying the other components like the RAM, motherboard, etc which might affect the computation speed for fluent?

Make sure you get a chipset that can use memory faster than 1600mhz. Memory speed and latency has more to do with computational speed on unstructured meshes than CPU frequency.

Thanks Kyle. I'll mostly go for a Gigabyte B75 mobo then. :)