Ryzen 3950x vs Ryzen 5950x

  • 14 December 2021
  • 3 replies
  • 314 views

Userlevel 1
Badge +1

Hi All,

I am currently using a Ryzen 3950x processor for NSC raytracing. I recently ran a trace of the same exact file on a 5950x processor and it was almost 10 times faster… All the other specs on the computers are the same.

The file involves some CAD, source files and scatter profiles. 

I tried to do a more simple speed test comparison by tracing some Zemax native sources through a single lens. When doing this, the computation time was about the same… 

This makes me think there is something bogging down the file on my computer, but I can’t figure out what it could be. When I share the file with the other computer, I am sending a .ZAR, we are both using the same version of Zemax and we aren’t running anything else on the computers.. Is there something else I should check for?

The other option is the 5950x is just that much better on complicated files… That is hard to believe..

Or maybe I should just use this as an excuse to upgrade to the Threadripper?

 

On a similar note, does anyone have any good benchmark tests for non-sequential to compare raytrace speed?

 

Thanks,

Kaleb 


3 replies

Userlevel 6
Badge +2

Hi Kaleb

We had a post on a similar subject here: 

Here is the CPU comparison: https://www.cpubenchmark.net/compare/AMD-Ryzen-9-5950X-vs-AMD-Ryzen-9-3950X/3862vs3598

Do you have a file you could share as we have an ongoing investigation from our developers?

Userlevel 1
Badge

Hi Kaleb,

You mentioned that when you tried more simple speed tests, the computation time was about the same. In my opinion, this is correct. Because you were testing “simple” tests, “simple” = less computation time. You have two PCs with different CPUs. One is faster. The other is slower. When the test file is simple = the computation time is shorter. The CPUs difference may be irrelevant since the task will finish quickly.

For the 10 times faster test between two CPUs with CAD, source files and scatter profiles, you probably have a complicated system and doing huge number of ray tracing. In this case, 5950X and 3950X does make a difference. 5950X has Zen 3 core Architecture and other key features that 3950X doesn’t have. There is no doubt that 5950X is faster. But does 10 time faster make sense? Could be. 

Do you mind letting me know if your two machines, 5950X and 3950X, both have the same software installed? both Windows updated to the same release?

Userlevel 1
Badge +1

HI Sandrine and Paulie,

Thank you for your replies. Unfortunately I cannot send the file over so I know that makes this pretty much impossible to troubleshoot…  This site is saying raytracing should be about 12% faster on the 5950: https://www.wepc.com/tips/amd-ryzen-5950x-vs-3950x/. That could be a good enough reason on its own to upgrade … 
Everything is the same between machines except for the CPU. 

Paulie - what you say makes sense about simple systems being hard to compare between computers. That’s why I was hoping someone has a good benchmark test that they run in non sequential mode to test out computer speeds.  

Since I’m not able to share the file, I don’t think we’ll be able to know exactly what’s going on, but I just wanted to put this out here in case anyone notices similar behavior.. 

Thanks!

Reply