PC is built: Help with benchmarks

[FutureLuxury]

I've built and setup my new (1st) PC and I'm wanting to run a few benchmarks to see if the system scores in the expected range.
CPU: Ryzen 5950x
GPU: MSI rtx 3090 Suprim X

 

I don't really know what I'm doing but I'm looking at running Cinebench + Furmark + Heaven.
Are there other tests I should run?

 

I tried doing an initial test (I have no overclock on the CPU or GPU):

 

- Furmark, 1080 preset for 1 minute
SCORE:21,007 points (350 FPS, 60000 ms)

 

- Cinebench R23
multi core: 26,386 pts
single core: 1,644 pts

 

Are these score good / okay for the components?

 - I'm also not sure what settings to use for Heaven.
 - Should I run a higher resolution for Furmark?
 - What should I use if I want to test temperatures under load?

 

Sorry for all the questions.

 

[SorryBella]

14 minutes ago, FutureLuxury said:

I've built and setup my new (1st) PC and I'm wanting to run a few benchmarks to see if the system scores in the expected range.
CPU: Ryzen 5950x
GPU: MSI rtx 3090 Suprim X

 

I don't really know what I'm doing but I'm looking at running Cinebench + Furmark + Heaven.
Are there other tests I should run?

 

I tried doing an initial test (I have no overclock on the CPU or GPU):

 

- Furmark, 1080 preset for 1 minute
SCORE:21,007 points (350 FPS, 60000 ms)

 

- Cinebench R23
multi core: 26,386 pts
single core: 1,644 pts

 

Are these score good / okay for the components?

 - I'm also not sure what settings to use for Heaven.
 - Should I run a higher resolution for Furmark?
 - What should I use if I want to test temperatures under load?

 

Sorry for all the questions.

 

Keep your text color to automatic. The people in dark mode cant read black text.

 

15 minutes ago, FutureLuxury said:

 - I'm also not sure what settings to use for Heaven.

You dont. You use the updated Unigine Superposition instead because a lot of game nowadays default to DX12 which Superposition always takes better advantage of.

 

17 minutes ago, FutureLuxury said:

 - What should I use if I want to test temperatures under load

Furmark and Prime95 is the classic free combos for reliable 100% load, but Aida64 is a better CPU stress tester. You can monitor temperature with slew of sensor readers but HWiNFO64 is my flavor for now.

[FutureLuxury]

27 minutes ago, SorryClaire said:

Keep your text color to automatic. The people in dark mode cant read black text.

Sorry, I didn't realize I'd changed the text from automatic. Maybe it happened in copying the body of text from a notepad?

27 minutes ago, SorryClaire said:

You dont. You use the updated Unigine Superposition instead because a lot of game nowadays default to DX12 which Superposition always takes better advantage of.

I'll do that instead, thank-you.

27 minutes ago, SorryClaire said:

Furmark and Prime95 is the classic free combos for reliable 100% load, but Aida64 is a better CPU stress tester. You can monitor temperature with slew of sensor readers but HWiNFO64 is my flavor for now.

In doing that, do I just run the test for an extended period of time and monitor that the temperatures stay in a safe range?

[FutureLuxury]

39 minutes ago, Caroline said:

What's the power supply? do a couple RealBench runs, you could also try OCCT which is basically Furmark + a CPU stress test, tbh you could do Furmark + IBT and the result would be the same, those are just to test the cooling and overclocking configs, since you're not overclocking the temps should stay within normal margins.

 

1080 is low for furmark, run it at native resolution, if native is 1080 I don't know what to say, a 3090 with a 1080p monitor? king of wasting money.

1 minute isn't going to test anything. Do an hour if you want to check temps/airflow but that's about it, synth benchmarks aren't great to measure real world performance, you gotta run 3D intensive programs for that, games, CAD, rendering, video processing and so on.

 

I only run benchmarks when I overclock something or want to see if a 20 year old PSU with half its components replaced blows up after an hour.

My monitor is 1440. I'll adjust. But even if it was 1080 I don't think it would necessarily be a waste in my case and I do not game and purchased it for 3d modelling, rendering and real-time 3d virtualization of large architectural files which are intensive regardless of my screen resolution and the performance will save me time when deadlines and time pressures are imposing.

I will be switching my CAD and rendering work to the PC in question from today so I guess I'll find out real world performance going forward.
I was just curious to get some comparative performance 'score' even if it's not entirely relevant.

[SorryBella]

19 minutes ago, FutureLuxury said:

In doing that, do I just run the test for an extended period of time and monitor that the temperatures stay in a safe range?

Yep. Also have a look at the speed. The clock speed under load should still be close to the boost clock on the box thanks to AMD's precision boost having no time limit.

[FutureLuxury]

16 minutes ago, SorryClaire said:

Yep. Also have a look at the speed. The clock speed under load should still be close to the boost clock on the box thanks to AMD's precision boost having no time limit.

Great, I'll do that. Thank-you again