Quad core or dual core for blender?

I am currently looking into buying a powerful desktop to replace my poor laptop that definately isn’t designed for 3D graphics or development but i manage simple things in blender.

I was wondering whether Blender prefers powerful dual cores or less powerful quad cores?
I will be mostly be using the BGE but also some rendering. I want to ba able to produce a decent game with a high frame rate (my laptop struggles to maintain 10fps on my awful first game)

Thankyou for all suggestions and help in advance and feel free to point me in the direction of new computers you may know about. :slight_smile:

The more cores you have the better it generally is for rendering. I’d recommend you do a forum search as this type of subject comes up pretty regularly

The more threads you can enable, the faster will be the rendering. I think its better to buy Quad Cores, since it helps in gaming too. I saw a great difference in GTA4 between an AMD Phenom X2 and X4.

Be certain that the computer you buy can actually use all of the cores that it has, and that it can actually do so continuously. Some lesser-quality motherboards are not speed-matched to the CPU as they ought to be, with the result that all of the cores cannot actually operate at their rated speed. You can also have serious heat-dissipation problems such that the CPUs have to be slowed-down, or stopped outright. “Sure, it’s nice that on-board heat sensors can do such a thing automatically, but the whole point of having such a system is that this kind of problem does not occur.”

Remember also that the behavior of a core is not exactly the same as that of a true second CPU. The cores are separate computational regions on a single silicon chip, but they are not each equipped with their own memory input/output circuitry. And it so happens that 3D rendering is a hugely memory-intensive as well as CPU-intensive task.

If you get something like an i7 from Intel, it will redirect power from unused cores to boost the performace of 1 or 2 cores. So in reality you get the best of both worlds.

When rendering you will want as many cores as possible, when modelling you will probably only care about 1 or 2 cores.

Intels chips are pretty decent.
The new i5 and i7 chips are based on the xeon architecture.
Having direct access to the memory is good.

The intel i5 seem to be equal to better than the new AMD x6 six core.
AMD seems to be a better price performance service in some articles
but when you get into very detailed test articles you will notice that they
basically flame it. Also the AMD chip is based on a quite out dated architecture.

The i7 is a good chip and you can get it at various price stages from 300 and

The i5 does not offer hyperthreading thus I would rather do with a cheaper i7.

But as mentioned make sure the motherboard and ram are making use of the

If money is tight well go with an i5 at the end they are all faster than your
current laptop anyway.

Prior to all, Quadcore, not Dualcore in answer to the OP.
Question is which one =)

Thats a bit far fetched =)

Xeons are based on the Core2 architecture on which the C2D and the C2Q families are based on.
The Core-i architecture is Nehalem based (which is distant related to the Core2 architecture) on which the Ci3, Ci5 and Ci7 families are based on.

More less yeah.

I took it up to convert benchmarkresults into % performance for a better overview.

Top of the list is the Ci7-980X (1000 Euro).
We take it, it performs 100% in Cinebench, x264 and Truecrypt (which are all processor heavy and multithreaded applications)

CB .....: Cinebench
TC .....: Truecrypt
pp%p ..: price per % performance

Processor          |  Price  |  CB  | x264 |  TC  | pp%p in CB
Ci7-980x           | 1000 Eu | 100% | 100% | 100% | 10,00 Euro
Phenom2 X6-1090T   |  300 Eu |  68% |  59% |   9% |  4,41 Euro
Ci7-920            |  220 Eu |  60% |  54% |  62% |  4,66 Euro
Phenom2 X6-1055T   |  200 Eu |  61% |  52% |  16% |  3,27 Euro
Ci5-750            |  170 Eu |  51% |  42% |  46% |  3,33 Euro
Phenom2 X4-965BE   |  150 Eu |  50% |  64% |  52% |  3,00 Euro
Phenom2 X4-925     |  120 Eu |  42% |  35% |  52% |  2,85 Euro

What also counts is the power consumption measured with Cinebench:

Normalized and ordered
Phenom2 X6-1090T   | 100% Performance / Watt
Ci7-980X           |  93% Performance / Watt
Phenom2 X4965BE    |  74% Performance / Watt

Unfortunately I got no power consumption data on the Ci5

For gaming the whole thing looks completely different, but the Ci5-750 is a very nice chip. I would guess it has a similar power consumption as the Ci7 so a slightly worse performance/watt but its overall performance, its price and power consumption makes it very intresting.

In direct comparison for applications I would choose a:
Ci5 over a Phenom2 X4 for its price/overallperformance and power consumption
Phenom2 X6 over a Ci7 for its price/renderperformance and power consumption if I had a AM2+/AM3 platform already.

For gaming and general system performance I would choose the Ci7 over the Phenom2-X6

But bottomline is, the more you spend, the more power you get.
And using a processor for work or professional you got to crunch numbers because then time is money and you got to see how fast the expenses make up for the time gained, or in other words, how much money you loose waiting because you saved on money for the hardware.

For “private” persons the difference is marginal i´d say and i would always recommend to make the budget choise.

hth and was not too technical =)

Thankyou everyone for contributing to this thread and helping me out.
Special thanks to arexma for brilliant information about different processors. :slight_smile:

I’ve currebtly been looking at the Dell XPS range (despite some people’s problems with Dell computers)
and am looking at the AMD phenom2 x6 for £698.99 with RAM upgraded to 6gb for an extra £40.

Or the Dell XPS 8100 for £729 with either a dual i5 or low end quad i5 that does not have hyper threading.

XEON and Core 2 Duo might have a realtionship
but I think we can agree on that there are great
differences between a core 2 duo chip and
a i5 / i7 chip.

it starts with the memory access, to core addressing
to, and ends in the performance gain.

the mobile i7 is 3 times faster than the mobile core 2 duo
I have.

That is quite a jump.


I was interested in trying to find a way to hook up 3 x6 on one board
to have a nice rig, because price wise that would be a great option.
Sadly that seems not to be possible.

The issue with some test reports you can read is that well you do not
know how comparable they are. Plus 3D, zip, and video tests actually
measure different aspects of you chip and rig.

But I question how much that is relevant for a user. As long as Blender
runs faster on chip and audio and video cutting is going well what else
do you want.

I would be careful with the AMD x6 since the 6 cores are equal
to slower than same price range i5 quads.

But at the end it also depends on what you can live with. If a rendering
is 15 minutes faster or slower but and it doesnt bother you while
saving money than I would go with that.

From the high end tests I read and talk to testers I cannot say that
anybody favored AMD when performance was a concern.

Xeon is no architecture, its a brand. Xeons where made from P6, NetBrust, Core and Nehalem, latter commonly known as core-i

The P6 and NetBurst architecture Xeons where pretty “unknown”, the rise began with the Xeon Core, which as the name says bases completly on the Core Duo architecture.
The Xeon Nehalem, Ci3, 5 and 7 all are based on the Nehalem architecture.

Was prolly my fault, the Xeon Core was the Xeon I assumed you meant the i5,i7 are based on the Xeon Core which is wrong.
So is that the i5,i7 are based on the Xeon Nehalem, as all those are based on the Nehalem architecture. =)

Basically a Xeon is just a derivate of the consumer processor with multiprocessor support but both are based upon the same architecture.

But yeah we can agree on the greatest difference between Core and Nehalem.
Nehalem finally has an integrated memorycontroller and you don´t have to buy a new mainboard for each processor because the MC was part of the SPP.

GSI Technology just introduced a 5 nanosecond SRAM, down from 9 nanosecond. So if my math is right, 200GHz processors are now enabled with 0 wait states. wowzer.

That is one of the nice things that the new chips are more plug n play than before.

It is impressive to see how much the i7 is getting close to the Xeon.
Previously the Xeon was a workstation rated chip. And you see PCs with i7 now
left and right.

Thanks again for all this infromation guys and I will hopefully be getting a new computer soon.
I have just been wondering whether less cores with higher clock speeds is better for modelling and genrally using blender rather than rendering. With my current laptop, Blender will only run on one core at 80% and 20% of the second core when modelling and stuff. It will only use both cores when rendering.
This may be due to the motherboard as mentioned by someone else.

I don’t know if there is a way to make Blender use more of the cores but if it will only use one then surely a high powerd dual core is better than a low quad?

Or am I missing something?
Thanks in advance.

Most stuff in blender is (still) dualthreaded some (still) singlethreaded, so a powerful dualcore is faster than a lower clocked quadcore. There are significant differences in the dual and quadcor architecture though that make the up for the lower clocks. For rendering it´s the more the merrier, the faster the better.

However, maintaining high framerates in your games in BGE depends on GPU power, not CPU.