Rendering is very slow with Intel Core i7

[Anonymous]

Hi, I have an inconvenience and hope someone can help me.

I have been working renders on a computer (described below) and it took about 8 minutes 40 seconds on a particular task:
Intel Core i5 3.4 GHz, two cores, fourth generation
8GB RAM
5400RPM HDD
No video card.

Now, I acquired new computer (described below) and doing the same task takes 6 minutes 20 seconds.
Core i7-6700 3.4GHz, 4 cores, 6 generation
16GB RAM
SSD 480
Video card NVIDIA GTX 970 4GB

That is, by a computer much higher, I only saves about 2 minutes.

I can do to reduce the time?
There will be incompatibility between Core i7 and 3ds Max 2016?

I have installed the operating system:
Windows 8.1 x64
3ds Max 2016
I am using Scannline for rendering.

Thank you very much for your support !!

[dgorsman]

How much less time were you expecting?

Render efficiency can be as much about the render settings and models as hardware.  Please provide more information, screenshots, etc. of what you are trying to do - maybe someone can help with tweaks to settings or constructive critique of what you are trying to render.

[Francisco_Penaloza]

From my understand scannline rendering is not that optimized for multi core CPU as any other bucket rendering engine.

Your new CPU has the same score speed than your i5 cpu. in reality the extra push in rendering you are getting my be the turbo boost that your new CPU has.

If you were using Mental Ray or V Ray you'll see more differences. but honestly your Fourth generation i5 was already a pretty good CPU.

If you really need to optimize renderings you may need an CPU that you can Over clock or get something in the real of 4Ghz for scannline rendering.

[Anonymous]

Thanks for the support dgorsman

As the core i5 dual-core and quad-core Core i7 , I thought it would halve the rendering time .

What settings should I do to optimize render time ??

I rendering image.

[Anonymous]

 Thank you for your contribution Francisco.Penaloza

I also think the processor speed influences the rendering time. I wish that other fellow think about it.

About "scanline" any of you could confirm if "scanline" can handle multicore and hyperthreading ??

What rendering engine do you recommend me to optimize the time?

Thank you very much for support me!!!

[PROH]

Hi. Yes the Scanline renderer can use multi cores, and I think also hyper threading. But that doesn't mean that it scales 100%. I don't know of any tests done with Scanline regarding this, but if the job you're talking about was the exact same on both i5 and i7, then it looks like it doesn't scale 100% (more like 25%).

Scanline isn't an especially fast renderer, so you could try using mental ray. I'm quite sure it scales better, and no matter what it'll give a far better result in most cases. Another possibility is to take a look at the Corona renderer. It scales almost 100%, and is very easy to use compared to most other advanced renderer. But it doesn't come free, like mental ray - you'll have to pay for it.

Hope it helps

[Anonymous]

Thank you for help me!!

If scannline supports multicore and hyperthreading, Core I5 has 4 thread and Core i7 has 8, the scale of speed should be 100%.

I think scannline supports multicore, but no hyperthreading, it is according to what is happening to me

What do you think??

Why do you think I'm going to have better performance with Corona renderer or mental ray??

Thank you very much for your support!!

[Francisco_Penaloza]

Scannline support multi cores and hyperthreading too, but compared to Bucket rendering it won't use all core in an optimized way.

Bucket rendering is a lot newer than Scannline method, that's why Bucket rendering is faster, the latest way to render in Progressive, some people proclaim it is faster than Bucket method, but it uses more memory for sure.

If you have any bucket rendering such, Mental Ray or V Ray, you'll get 100% use of all your cores.

For instance, if you have one core as base, two cores will render double speed, then tree cores will render tree times faster and so on, until one limit where render speed/ cores amount hit a plateau.  Then distribute rendering is better.