Jump to content

Understanding Ryzen 3900X benchmarks compared to i7 8700K


Recommended Posts

I'm considering building a new system, as my main work machine is currently a laptop. I was going to go with the AMD Ryzen 3900K processor for handling very heavy scenes, optimizing more for test renders over final renders, but still doing some very large final renders occasionally.

 

I decided to benchmark it against the Intel processor in my laptop, an Intel i7 8700K, and surprisingly, the AMD is only 2-7% faster depending on core use (though it didn't show all 12). I was expecting it to be a lot higher. I would want at least a 25% increase to consider it worth the hassle and cost of building a new system. I'm assuming that system doesn't pay attention to core count or something?

 

Yet when I go to ChaosGroup's benchmarks (I use V-ray), the Ryzen 3900X gets almost double the score.

 

So are sites like UserBenchmark at all useful for the type of work we do? As I understand, the ChaosGroup benchmark only is about final renders, and I don't know where to go to look for modeling with heavy geometry/scenes.

 

I also assume that software like 3DS Max is optimized for the Intel chipset. Is the 3900X still the way to go compared to something like the i9 9900k, if my primary goal is to reduce lag and secondary is to speed up test renders? Does Turboboost work similarly for either for the work we do?

 

Thanks.

Edited by kirstenzirngibl
Link to comment
Share on other sites

If you are doing test renders frequently,than AMD is the way to go, and for every other purpose, AMD is still the way to go.

...

For viewport manipulation, single core or turbo is responsible, and in that segment AMD is a little bit slower than Intel, but if you have some good graphic card, 1080Ti +, you will notice no differences.

...

So, YES, go with AMD.

Link to comment
Share on other sites

Kirsten,

 

You can also try checking on Corona's results:

https://corona-renderer.com/benchmark/results

 

Many of 3DS Max's functions are single-threaded so you still need to consider the single-core speed for working in the viewport and using tools like conform for example.

 

I'm curious what impact if any the Intel Embree raytracing library has on an AMD vs Intel for rendering.

Link to comment
Share on other sites

Synthetic benchmark needs to be taken with a pint of salt all the time.

There are many type of benchmarks around, they all test different functions of the CPU. and also the system that are tested on, will affect the performance of those test.

 

 

You need to consider also that heat is an important issue on laptops.

All laptops when get hot, the performance is reduced to protect the system.

Desktop computer are not affected much for this heat issue and you can do work around easily and cheap.

 

 

Considering all that, if you will be doing arch viz in your machine, then you should look to 'real life' test, such Corona render test, Blender render test, Chaos group render test, even Cinebench is closer to what we do than a hand break or Win Zip test. I don't understand why they still using POVRay to test ray tracing.....LOL

 

 

At my office I have a Xeon CPU, 6 cores running at 3.6Ghz Turbo to 4.0Ghz. at Home I have a Ryzen second get 8 cores 16 threads OC to 3.9 Ghz.

My home machine is considerable faster than my office work station overall, test renderings and final renders are done quickly compared to a less number of cores CPU running about the same speed.

 

Yes 3D Max is mostly single core for mostly all modeling operations, in that case the faster your nominal cpu speed is, the better, that's why I did Over clock my Ryzen CPU to have a better performance while modeling.

 

 

Lots of Ram and a decent video card will help too.

Now regarding video cards, and with not intention to start the ethereal battle of GTX vs Quadros, at home I have a Quadro P4000; at the office I have a RTX 2080Ti, honestly I get more smooth overall view port performance with the Quadro video card than the RTX I feel very desapointed in that seance. Of course when I do GPU rendering then the RTX shows it place as a top end card, but for Modeling in 3D Max, REVIT, Cinema 4D, MODO. The Quadro performs the same if not better.

3D Max has become very unreliable with RTX or GTX card, there are many glitches that you don't get with Quadro cards. Also they discontinued the support for any 'gaming' card, if there was an support previewsly any ways, I gues now is official. :p

 

 

But again, if you use GPU rendering or other CUDA necessary task then the more CUDA cores the better.

My two cents.

Link to comment
Share on other sites

It's also worth noting that the site Userbenchmark has a bias towards intel processors, they changed the way they calculate scores earlier this year. Here's an example.

 

You can trust Chaosgroups benchmarks. The 3900X has 12 Cores/24 Threads and the 8700k has 6 Cores/12 Threads at similar speeds. This lines up with Chaosgroups benchmark of about double the speed. A 3900x will absolutely tear up renderings!

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...