Hi, I am running the 4K benchmark to determine overhead for some new systems going into a permanent install.
I am running all 20 noise layers and outputting the source to NDI to stress the system. My goal is to get 4K 60fps while outputting to NDI Studio Monitor.
I am testing on 3 systems.
System 1 - 10 core Intel Xeon-W 2255 base clock 3.7Ghz 64GB ram + RTX A4000 GPU
System 2 - 14 core Intel I-9 10940X base clock 3.3Ghz 128GB ram + RTX A5000 GPU
System 3 - 16 core Ryzen 5950X base clock 3.4Ghz 64GB ram + RTX A4000 GPU
When I run the benchmark and just select all 20 noise layers
System 1 - 60 FPS
System 2 - 60 FPS
System 3 - 48-52 FPS
When I run the benchmark and just select all 20 noise layers with NDI Monitor
System 1 - 40 FPS
System 2 - 45 FPS
System 3 - 30 FPS
Why is the Ryzen not performing as well as the 14 core Intel?
Also does anyone have a 24 core Threadripper who would benchmark this for us?
AMD vs Intel Benchmark
Re: AMD vs Intel Benchmark
I have a few 3960x builds.
Is this just for NDI testing? DXV3 uses your GPU.
What do you need tested? I might have a few free min today to do that. How many external displays do you need tested? I have (3) 3G displays connected via 3080ti
Is this just for NDI testing? DXV3 uses your GPU.
What do you need tested? I might have a few free min today to do that. How many external displays do you need tested? I have (3) 3G displays connected via 3080ti
Re: AMD vs Intel Benchmark
I tried to direct message you if you did not get it let me know.Arvol wrote: Thu Aug 25, 2022 15:13 I have a few 3960x builds.
Is this just for NDI testing? DXV3 uses your GPU.
What do you need tested? I might have a few free min today to do that. How many external displays do you need tested? I have (3) 3G displays connected via 3080ti
I basically am trying to test the 4K benchmark from Resolume with all 20 noise layers active.
Then set the output to NDI and view in studio monitor on the same system/display.
The FPS should be at 60fps before the NDI Studio Monitor. I want to know the fps once the NDI monitor is running.
Thanks.
Re: AMD vs Intel Benchmark
For anyone else viewing this thread. I know I only asked for someone with a Threadripper but if someone has a 24core Intel that would be great for running this quick test as well.
Re: AMD vs Intel Benchmark
20 you have 20 layers of noise running at once (That is 100% GPU load). The A4000 isn't going to perform as good as the others.
Do you just have 1 NDI stream outputting? (CPU load). If it's just 1 NDI stream, you can do that from an i5. If you have 20 NDI streams running.... then you'll need a beefy CPU.
Most of what Resolume does is GPU based. The only time I tell people to focus on CPU is when they are dealing multiple NDI steam (4+) or multiple applications running at once. The priority should be GPU on these builds. System memory will come into play with deck load times (thumbnail loading) and certain FX. (Delays for example)
Do you just have 1 NDI stream outputting? (CPU load). If it's just 1 NDI stream, you can do that from an i5. If you have 20 NDI streams running.... then you'll need a beefy CPU.
Most of what Resolume does is GPU based. The only time I tell people to focus on CPU is when they are dealing multiple NDI steam (4+) or multiple applications running at once. The priority should be GPU on these builds. System memory will come into play with deck load times (thumbnail loading) and certain FX. (Delays for example)
Re: AMD vs Intel Benchmark
I am running all of Column2 and just outputting the Comp to one NDI Studio Monitor. Its a personal stress test for a project that I do not have the graphics for yet so just trying to get and idea of over-perfromance.Arvol wrote: Fri Aug 26, 2022 19:08 20 you have 20 layers of noise running at once (That is 100% GPU load). The A4000 isn't going to perform as good as the others.
Do you just have 1 NDI stream outputting? (CPU load). If it's just 1 NDI stream, you can do that from an i5. If you have 20 NDI streams running.... then you'll need a beefy CPU.
Most of what Resolume does is GPU based. The only time I tell people to focus on CPU is when they are dealing multiple NDI steam (4+) or multiple applications running at once. The priority should be GPU on these builds. System memory will come into play with deck load times (thumbnail loading) and certain FX. (Delays for example)
Here is a link to view what I did: https://replay.dropbox.com/share/RtIArGeOUMAvhGET
Re: AMD vs Intel Benchmark
Also the GPU on this system is an A4000 as well.mronis wrote: Fri Aug 26, 2022 20:34I am running all of Column2 and just outputting the Comp to one NDI Studio Monitor. Its a personal stress test for a project that I do not have the graphics for yet so just trying to get and idea of over-perfromance.Arvol wrote: Fri Aug 26, 2022 19:08 20 you have 20 layers of noise running at once (That is 100% GPU load). The A4000 isn't going to perform as good as the others.
Do you just have 1 NDI stream outputting? (CPU load). If it's just 1 NDI stream, you can do that from an i5. If you have 20 NDI streams running.... then you'll need a beefy CPU.
Most of what Resolume does is GPU based. The only time I tell people to focus on CPU is when they are dealing multiple NDI steam (4+) or multiple applications running at once. The priority should be GPU on these builds. System memory will come into play with deck load times (thumbnail loading) and certain FX. (Delays for example)
Here is a link to view what I did: https://replay.dropbox.com/share/RtIArGeOUMAvhGET
Re: AMD vs Intel Benchmark
In my experience Base Clock Speed > Core Count. I've had good results on quad core systems in the past.
-
- Posts: 2
- Joined: Thu Sep 01, 2022 20:06
Re: AMD vs Intel Benchmark
Anche se sembra strano che i problemi li hai solo con AMD ma, la rete è una 10gb? Hai provato a mandare al solo pc AMD?