Ideal computer specs for VRED rendering?

[KD637]

It’s time to upgrade from my 9 yr old MacBook Pro. Tired of leaving it on overnight just to render a medium quality image. I’m switching to a PC desktop since Autodesk is discontinuing Alias and VRED for Mac. 

I found what looks to be a pretty good deal for a desktop PC. If any tech geeks and/or car modelers would like to chime in, please let me know if these specs are adequate for current and near-future versions of VRED and Alias.

Here are the specs:

• Intel Core i5+ 8400 (2.8 GHz, 6-Core)
• 16GB RAM, DDR4-2666 (upgradable to 64GB)
• NVIDIA GeForce GTX 1060 (6GB GDDR5, VR-Ready)
• 250GB SSD, 1TB HDD

How well do these specs hold up for present and near-future versions of VRED and Alias? I don’t do anything crazy like video editing or special effects, but I do like to make some photo quality renderings and simple animations (camera panning, doors opening and closing, seats moving, etc.)

Any feedback is much appreciated!

[Christian_Garimberti · Accepted solution]

Hi,
I think that the i5 will be a bottleneck. With only 2.8 ghz...
While you are in open gl you’ll use one core at a time, so 2.8 instead a 3.6 ghz i7 will be 40% slower.
And the same for raytracing. You’ll use all the 6 cores, but all the 6 cores (12 if you activate HT) will work at 2.8.
I’ll check for a faster CPU. Also, fastest is the cpu, better the GPU works.

Chris

[seiferp]

Just want to point out again that the GTX cards do work and probably most of you are using them... but they are not officially supported!

[KD637]

---

@Christian_Garimberti wrote:
Hi,
I think that the i5 will be a bottleneck. With only 2.8 ghz...
While you are in open gl you’ll use one core at a time, so 2.8 instead a 3.6 ghz i7 will be 40% slower.
And the same for raytracing. You’ll use all the 6 cores, but all the 6 cores (12 if you activate HT) will work at 2.8.
I’ll check for a faster CPU. Also, fastest is the cpu, better the GPU works.

Chris

---

Thank you for the info. So Open GL only uses ONE core at a time, even though CPUs have multiple cores? That's good to know. 

Besides the CPU, do you think the GPU and RAM can hold up?

[KD637]

---

@seiferp wrote:

Just want to point out again that the GTX cards do work and probably most of you are using them... but they are not officially supported!

---

Hi, can you please elaborate? I notice on the website Autodesk only recommends the NVIDIA Quadros. Is there a difference in performance between those and the GTX series? Thanks!

[Christian_Garimberti · Accepted solution]

Maybe i said something wrong...

also in OGL Vred seams to use all the cores...

the peaks are when i rotate the scene.

But, the speed of the cores remains very important.

About the difference between GTX and Quadro:

Gtx is for gaming, Quadro for intensive CAD use.

More or less the architecture is the same, but if you use the GPU for professional use, Quadro is the better choice.

Quadro Cards are made to work under stress for hours and hours without problems.

But GTX is very cheaper.... 

Chris

[cikho]

another big difference between the cards are drivers.

quadro drivers are certified to work with cad products, hence cad developers use these to create their softwares..

what can occur with gaming drivers is that some optimization made for games or other applications makes your cad crash unexpectedly.

this for example happened for a certain range of gaming drivers with showcase back in 2013.

having said this, gaming hardware still offers powerful solutions at accessible prices, and might actually have improved performance: given a finite budget, a low end quadro might cost more than a high end geforce that is actually performing better..

[michael_nikelsky]

In OpenGL rendering there are parts that are multithreaded and use multiple cores like view frustum culling for example. But the rendering part itself, so sending the workload to the GPU, only happens on one Core. There will still always 2 Cores used in that case as well since the driver also runs in a separate thread and transforms the draw commands we send to a form that the GPU understands. But in general a CPU with few cores but high clock speed will perform better in OpenGL than one with many cores and low clock speed. 

For raytracing it is completely different since raytracing is highly parallel and therefore can make good use of many cores.

Kind regards

Michael