Underperforming RX 5700 XT

[Praetvereaucrat]

Hi everyone,

 

I've recently built my first PC, and as new builders are, was very excited to test out the new system. I ran a benchmark in UserBenchmark, and my GPU performance came out to be way below expectations. This is also after updating drivers for the GPU. I've attached a screenshot of my benchmark to this post. Any feedback/troubleshooting is appreciated!

 

PS I start classes in a bit, so I'll try to respond with additional details ASAP after I'm done with my classes today. Thanks again

 

EDIT 1:

Hi everyone,

 

Thank you for all your replies! Yes, I did run UserBenchmark after a fresh install. I just did another benchmark with Superposition as well, as some of you suggested, and attached the results to this post. The reflected performance does seem a bit lower than expected it seems for SP too. I'm also looking at 3DMark to use as a benchmark as many of you had highly advised. If y'all could provide additional feedback/troubleshooting, it would be greatly appreciated once again!

 

Cheers!

Edited September 12, 2019 by Praetvereaucrat

[Rainbow Dash]

Userbenchmark is not a really accurate way to test your system performance, due to different setups and hardware. Try running a GPU benchmark like 3DMark.

[Turtle Rig]

How is gaming performance ?  After all that is what matters.  What settings do you use in game and AMD settings ?  Is this after a fresh install ?

[Zando_]

Just now, Rainbow Dash said:

Userbenchmark is not a really accurate way to test your system performance, due to different setups and hardware. Try running a GPU benchmark like 3DMark.

^^^ Userbenchmark is terrible for actually benchmarking (it gives kinda really loose estimate-ish numbers that are sorta maybe alright and possibly ok), 3DMark has much better and more consistent benchmarks, TimeSpy would be the best one to run, FireStrike after that. Then you can look up people with the same CPU/GPU combo you have using advanced search and compare your scores (keep in mind the top scores are usually madlad overclockers so your score will be lower than those). 

[5x5]

Uaerbenchmark is a HORRIBLE source. Don't trust anything it says. They've be paid.off and are currently attacking tech YouTubers. 

[Stormseeker9]

Userbenchmark is absolute rubbish. Don’t pay any attention to that.

 

Run something like Heaven benchmark / Firestrike / timespy and compare results with other results.

 

use some in game benchmarks if available to you.

[Zando_]

Just now, 5x5 said:

Uaerbenchmark is a HORRIBLE source. Don't trust anything it says. They've be paid.off and are currently attacking tech YouTubers. 

Lol really? They attacked guys doing actual benchmarks or something? That never ends well, wonder who's managing that effort... 

[5x5]

1 minute ago, Zando Bob said:

Lol really? They attacked guys doing actual benchmarks or something? That never ends well, wonder who's managing that effort... 

Look at their about page. They're openly attacking HU and open to a defamation lawsuit. Hope their pockets are deep

[Zando_]

1 minute ago, 5x5 said:

Look at their about page. They're openly attacking HU and open to a defamation lawsuit. Hope their pockets are deep

 

 Looooool. Tho tbh if I were a fish I'd buy hella water, better than suffocating. 

 

4 minutes ago, Stormseeker9 said:

Userbenchmark is absolute rubbish. Don’t pay any attention to that.

 

Run something like Heaven benchmark / Firestrike / timespy and compare results with other results.

 

use some in game benchmarks if available to you.

Heaven is a bit old, good for stress testing but Superposition is much better for comparing benchmark scores. 

[Stormseeker9]

31 minutes ago, Zando Bob said:

 

Heaven is a bit old, good for stress testing but Superposition is much better for comparing benchmark scores. 

old but gold  

[Zando_]

1 minute ago, Stormseeker9 said:

old but gold  

Doesn't push some newer GPUs as hard as something like Superposition 1080p Extreme (good gosh that benchmark brings any GPU to its knees), which will leave more room for inconsistent numbers. 

For stress testing it's still the best thing out there, Superposition throws fits if I put it on a loop but don't have it selected as the active window, whereas Valley has 0 issues if I run it behind my Afterburner/Wattman window as I change settings.

[Stormseeker9]

2 minutes ago, Zando Bob said:

Doesn't push some newer GPUs as hard as something like Superposition 1080p Extreme (good gosh that benchmark brings any GPU to its knees), which will leave more room for inconsistent numbers. 

For stress testing it's still the best thing out there, Superposition throws fits if I put it on a loop but don't have it selected as the active window, whereas Valley has 0 issues if I run it behind my Afterburner/Wattman window as I change settings.

okay, it was more of a pun than anything else  

 

DO agree on SP. But for me 3Dmark will always be the most reliable bench

[Zando_]

Just now, Stormseeker9 said:

okay, it was more of a pun than anything else  

 

DO agree on SP. But for me 3Dmark will always be the most reliable bench

*beams in having scored the 4th highest TimeSpy score for my CPU/GPU combo even though it's not a very common setup*

(5820K and Radeon VII if you don't wanna check my sig)