[SOLVED] 3080 underperforming

[Ranshii]

I just recently built a new PC with a 3080 FE, but when playing games and running benchmarks it seems to be falling behind benchmarks online. I ran time spy and got a GPU score of 15,993 and I also ran Port Royal and got a score of 10,413 both of which seem on the low end for a 3080. I've tried to reinstall drivers, made sure the RAM was at full speed, tried to overclock but got no real gain, is there anything I can do or am I just unlucky with my card?

Port Royal: https://www.3dmark.com/pr/509317
Time Spy: https://www.3dmark.com/spy/15351027


My specs are:
3080 FE
Ryzen 3700x
16GB DDR4 3600mhz ram
MSI B550 Tomahawk Mag

 

[MyMindIsBlank]

Probably a CPU bottleneck; an 8 core Ryzen 7 3800x or 5800x would be more ideal than a 6 core Ryzen 5 3700x.

 

[BradHP]

Probably a CPU bottleneck; an 8 core Ryzen 7 3800x or 5800x would be more ideal than a 6 core Ryzen 5 3700x.

3700x is the 8 core Ryzen 7. Should be fine paired with a 3080.

 

[Zerk2012]

Your probably on the lower end because it's a FE edition and not a aftermarket overclocked model.

 

[BradHP]

It shows that you're higher than 94% of other results on Port Royal, and higher than 98% on Time Spy. That seems pretty good. What are you comparing it to?

Most reviewers have it paired with an i9 or Ryzen 9 for testing, so they're going to get higher marks.

 

[MyMindIsBlank]

3700x is the 8 core Ryzen 7. Should be fine paired with a 3080.

Oh, duh. I feel stupid. I misread it as a 3600x.

 

[BradHP]

Oh, duh. I feel stupid. I misread it as a 3600x.

Well, your mind is blank, so it's okay.

 

[sizzling]

Ok I just ran these on my rig which is a 3700X (PBO on) and Gigabyte 3080 Gaming OC (no manual overclock) with 2x16gb 3600mhz CL16. 3DMark did an update just beforehand so is completely up to date and running 1440p.

Time Spy = 16,011
Port Royal = 11,504

 

[sizzling]

Noticed my average gpu temperatures 67c and temperature has a noticeable impact on boosting.

 

[BradHP]

Ok I just ran these on my rig which is a 3700X (PBO on) and Gigabyte 3080 Gaming OC (no manual overclock) with 2x16gb 3600mhz CL16. 3DMark did an update just beforehand so is completely up to date and running 1440p.

Time Spy = 16,011
Port Royal = 11,504

OP didn't say what resolution he's running at.

 

[sizzling]

OP didn't say what resolution he's running at.

I know, but I run 1440p and it provides a point of reference of a very similar setup.

 

[BradHP]

I know, but I run 1440p and it provides a point of reference of a very similar setup.

Yeah, that wasn't a hit against you. I meant that maybe he's running a 4k monitor or something and that's why his scores
are lower. If he's at the same resolution or lower, then he might have some other issues.

 

[Ranshii]

Your probably on the lower end because it's a FE edition and not a aftermarket overclocked model.

It shows that you're higher than 94% of other results on Port Royal, and higher than 98% on Time Spy. That seems pretty good. What are you comparing it to?

Most reviewers have it paired with an i9 or Ryzen 9 for testing, so they're going to get higher marks.

I've been basing it off graphics scores people have been receiving in both Port Royal and Time spy, which I fall short of, i keep seeing FE cards getting 17.5-18k on PR but I struggle to hit just 16k and on benchmarks where people use a similar system, I can fall short by 15-20 FPS

Yeah, that wasn't a hit against you. I meant that maybe he's running a 4k monitor or something and that's why his scores
are lower. If he's at the same resolution or lower, then he might have some other issues.

I use a 1440p 144hz monitor, should have mentioned that earlier

The GPU does run hot, when playing intensive games it'll usually sit at 75c, but even when I turn fans to 100% and even fully open the case and get the temps down to the low 60s there is no change in performance.

 

[Zerk2012]

I've been basing it off graphics scores people have been receiving in both Port Royal and Time spy, which I fall short of, i keep seeing FE cards getting 17.5-18k on PR but I struggle to hit just 16k and on benchmarks where people use a similar system, I can fall short by 15-20 FPS



I use a 1440p 144hz monitor, should have mentioned that earlier

The GPU does run hot, when playing intensive games it'll usually sit at 75c, but even when I turn fans to 100% and even fully open the case and get the temps down to the low 60s there is no change in performance.

They could have their card overclocked and you don't so it would still say FE card but it's running faster than yours.

EDIT they could of also been running the original driver for the card if I remember right the latest driver gimped the card a bit but made it more stable and stopped the shutdowns.

It's hard to compare apples to apples, you can buy 2 identical cars and one will outrun the other.

75C is not really hot for a video card.

Time Spy average score 15762 for a 3080 so your a bit above average!

Highest score: 21177
Average score: 15762
I see nowhere that it says if it was a FE Edition card.

https://www.3dmark.com/search?_ga=2.148902718.941436202.1605654148-424608440.1605654148

 

[BradHP]

I guess the real question at the end of the day is--can the card play the games you want to play at the fps you want to get? Benchmarks are fun, but have lots of variables to the numbers and are ultimately meaningless. Play some games and have fun.

 

[Ranshii]

They could have their card overclocked and you don't so it would still say FE card but it's running faster than yours.

EDIT they could of also been running the original driver for the card if I remember right the latest driver gimped the card a bit but made it more stable and stopped the shutdowns.

It's hard to compare apples to apples, you can buy 2 identical cars and one will outrun the other.

75C is not really hot for a video card.

Time Spy average score 15762 for a 3080 so your a bit above average!

Highest score: 21177
Average score: 15762
I see nowhere that it says if it was a FE Edition card.

https://www.3dmark.com/search?_ga=2.148902718.941436202.1605654148-424608440.1605654148

So 3dmark doesn't say whether it's an FE or from a brand like ASUS or MSI, so I'm going off the average, which I know isn't the best as that's going to include manually overclocked cards aswell as partner cards that have a factory OC, so i'd expect to be running a bit slower than that. But i'm also going off of reviews like for example here: https://www.guru3d.com/articles_pages/geforce_rtx_3080_founder_review,27.html which got 17.8k for a graphics score.

I guess the real question at the end of the day is--can the card play the games you want to play at the fps you want to get? Benchmarks are fun, but have lots of variables to the numbers and are ultimately meaningless. Play some games and have fun.

Don't get me wrong, even if I got a slower running GPU, it's still better than the vast majority of cards out there and is more than enough to get great framerates on games on max or very close to max without any issue. I'm still very happy with it, and I am lucky to even have one.

 

[Zerk2012]

So 3dmark doesn't say whether it's an FE or from a brand like ASUS or MSI, so I'm going off the average, which I know isn't the best as that's going to include manually overclocked cards aswell as partner cards that have a factory OC, so i'd expect to be running a bit slower than that. But i'm also going off of reviews like for example here: https://www.guru3d.com/articles_pages/geforce_rtx_3080_founder_review,27.html which got 17.8k for a graphics score.




Don't get me wrong, even if I got a slower running GPU, it's still better than the vast majority of cards out there and is more than enough to get great framerates on games on max or very close to max without any issue. I'm still very happy with it, and I am lucky to even have one.

From your own link they tested one card. If you look at my link their a bunch of user tests and your just above average.

I'm not sure what answer your looking for, you got a garbage card send it back?

1 Sample test as in your link is just one sample.

You test 30 cards you get a better baseline, you test 1,000 cards you get a even better baseline.

EDIT They don't list the driver version their using in your link are you using the same driver? That can change the score also.
Like I said above I think the latest driver gimped the cards a bit because people were having crashing problems from voltage spikes.