Is my 3080 underperforming?

Joined
Nov 21, 2020
Messages
8
So i recently got my hands on a RTX 3080 Asus tuf non-oced. And i proceded to bench it right away. Using Userbenchmark, i got a good average result on everything. Everything seemed to work fine, until i started to play watch dogs legion. I wanted to compare my gpu to this article (https://www.legitreviews.com/watch-...mance-benchmarked-on-12-graphics-cards_223425). According to this article, i should be able to hit 70 fps average on 4k very high settings/no rt/no dlss. I got only 58 average fps on my system. Using rivatuna, i can see that my cpu is only being 40-50% utilized, while my gpu is being 97% utilized. Therefore, i don't think it can be a cpu bottleneck. I have the latest Nvidia drivers (457.30 game ready drivers).
Please note that on 1080p ultra settings, my CPU is actually bottlenecking my GPU with a usage of 80% and a gpu usage at 85%
I am not too great at computers and need help here.

PC:
CPU= INTEL I7-9700k/no oc (can this really be holding my 3080 back?)
GPU=NVIDIA RTX 3080 Asus TUF/no oc
Motherboard=Z390-A PRO (maybe pcie 4 performs a lot worse on a pcie 3 mb?)
Ram=16gb 3600mhz xmp profile.
PSU=750 watt (maybe the problem Lies here?)
Monitor=BenQ XL2420T

Ps. My monitor has a native 1080p, so i am using DSR to increase it to 4K. Does DSR strain the GPU further than native 4K?
My temps are all around fine with 51 degrees on my GPU at 97% load and 61 degrees on my cpu at 80% load. All watch dogs legion tests were done with vsync off ofcourse.

Can it really be the difference between TUF and ROG strix? 12 fps? I assume the card the article used was overclocked aswell. I am also Playing on ubisoft+ does internet speed have anything to do here?

pps. I just did a 3dmark free edition test on and my GPU scores around 16700 i know that is fine for a 3080 but if the problem isnt my GPU, what Can it be? my Cpu? or maybe my PSU?
Thank you :)
 

madpistol

Limp Gawd
Joined
May 17, 2020
Messages
369
You said you got a 3Dmark score of 16700. What benchmark did you run (Firestrike, Time spy, etc.) and what was your CPU and graphics scores?
 

Damar

Supreme [H]ardness
Joined
Jun 20, 2004
Messages
4,596
My EVGA 3080 FTW Timespy benched just slightly less than the 2 x Titan Xp FE's it replaced at 3440x1440 at stock out of the box settings. Granted it ran a hell of a lot quieter and with less power draw. The 3080 didn't break 65C which to me was pretty great.

2 x Titan Xp's were 18,695 Graphics and 7,517 CPU overall of 15,285

EVGA 3080 FTW was 17,932 Graphics and 7,669 CPU overall of 14,934

Once I upgrade the CPU and rest of the system things should be a bit better, though I'm a little surprised at the initial numbers.
 

Mchart

2[H]4U
Joined
Aug 7, 2004
Messages
4,054
It won't be a lot, but the 9700k will give a little lower frames in a game like Legion, or any of the newer open world titles like AC Valhalla, etc. More threads helps with these games. It's not huge, but it helps. Is it 12 FPS average? Likely not, but between comparing your CPU to one with more cores and threads and comparing to an OC'd card that might be the difference.
 
Joined
Nov 21, 2020
Messages
8
My card runs at 1900 MHz when im playing. The Oced Cards run at 2000-2050 MHz. That together with my CPU might be the cause. I Think you are right Mchart. I hope you are. Would you guys be worried with my described performance plus This performance? -()
(Forgot to turn off my mic)
I know watch dogs is bandlyst optimized, but i am also falling 10 fps behind rdr2 Maxed out 4K benchmarks :/

Btw, notice the likely CPU bottleneck at 1080p?
Thank you for actually takling the time to comment on This. Means a lot :)
 

Mchart

2[H]4U
Joined
Aug 7, 2004
Messages
4,054
I don't have much more intelligent to offer, other than my own experience going from a 5820k to a 2700X build system (and eventually a 3800X). The 2700X compared to an overclocked 5820k should have been worse for gaming, but the extra cores/threads made a pretty noticeable difference in games like AC Odyssey, etc at the time. So comparing a non-SMT intel chip to a generation newer SMT chip that also has two more physical cores? It's probably a big difference.
 
Joined
Nov 21, 2020
Messages
8
My EVGA 3080 FTW Timespy benched just slightly less than the 2 x Titan Xp FE's it replaced at 3440x1440 at stock out of the box settings. Granted it ran a hell of a lot quieter and with less power draw. The 3080 didn't break 65C which to me was pretty great.

2 x Titan Xp's were 18,695 Graphics and 7,517 CPU overall of 15,285

EVGA 3080 FTW was 17,932 Graphics and 7,669 CPU overall of 14,934

Once I upgrade the CPU and rest of the system things should be a bit better, though I'm a little surprised at the initial numbers.
What clock is it running at? And is it oced?
 
Joined
Nov 21, 2020
Messages
8
Update: Just did a 3dmark test with this result.
All non-essential background programs were turned off.
Graphics score was around 6-700 points lower when i diden't use afterburner to crank up power +17% (can't touch voltage in afterburner). 17%+ power limit was the only thing i did in afterburner.

I don't know why it diden't validate? i tried this 2 times with simillar results. Yesterday it did validate.
 

Attachments

  • 3dmark.PNG
    3dmark.PNG
    2 MB · Views: 0

madpistol

Limp Gawd
Joined
May 17, 2020
Messages
369
Update: Just did a 3dmark test with this result.
All non-essential background programs were turned off.
Graphics score was around 6-700 points lower when i diden't use afterburner to crank up power +17% (can't touch voltage in afterburner). 17%+ power limit was the only thing i did in afterburner.

I don't know why it diden't validate? i tried this 2 times with simillar results. Yesterday it did validate.
Just did TimeSpy on my 3080, and it was literally identical to your score (1 point difference).

Your score is fine.
 

Damar

Supreme [H]ardness
Joined
Jun 20, 2004
Messages
4,596
What clock is it running at? And is it oced?
Stock out of the box. Boosts to 1950 solid and stays there. Temps rarely break 65C except in a few cases. I haven't pushed it yet really.

I guess I expected a bit more oomph, though it's not a slouch either since it's basically matching the performance I had with 2 x Titan Xp's in SLI in the games I played which used it well, and running cooler and much quieter.

FFXIV gets a bit more performance than before at higher detail levels.

I wish I had the money to drop for a better CPU right now but alas, I was lucky to get the 3080, nevermind a new CPU/MB/RAM like I want to.
 
Joined
Nov 21, 2020
Messages
8
Stock out of the box. Boosts to 1950 solid and stays there. Temps rarely break 65C except in a few cases. I haven't pushed it yet really.

I guess I expected a bit more oomph, though it's not a slouch either since it's basically matching the performance I had with 2 x Titan Xp's in SLI in the games I played which used it well, and running cooler and much quieter.

FFXIV gets a bit more performance than before at higher detail levels.

I wish I had the money to drop for a better CPU right now but alas, I was lucky to get the 3080, nevermind a new CPU/MB/RAM like I want to.
i still feel 16900 graphics score compared to yours, is a little low. Unless 50Mhz really does do such a big deal? maybe there is a difference because of different brands?

Temps on the 3080 is really great. I am absolutely stunned. Mine doesn't break 65C either. :)
 
  • Like
Reactions: Damar
like this
Joined
Nov 21, 2020
Messages
8
Does any of you know of a "time spy" graphics score from an Asus 3080 Tuf non-oc? I can't find any from the same card to compare.
 
Top