• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

LOW TIMESPY SCORES FOR 3080. HELP!

ikinson

ikinson

ikinson

Associate
Joined
3 Feb 2021
Posts
105
Hello, I’m new to this forum. Hope I’m posting in the right category.
So as the title says, I’m getting lesser than average scores on my 3080 fe paired with a 5600x. Users and reviewers got a graphics score of 17800-18000 at stock. I don’t like to fiddle around much expect for the fan curve so I have the default msi fan curve(1:1) applied and everything else at stock. Cpu is at stock too and only XMP is enabled. So my scores were as follows with Gsync disabled and balanced power plan:
dPt0JCL.jpg

Ultimate power plan on windows didn’t make much of a difference for me and my initial scores were only 17100-17200 with Gsync enabled. I read that my Gsync should be disabled for benchmarks and so I did and got the scores above.
My gpu reaches a max of only 65-66 degrees ingame which is way cooler than most cards out there. I’m aware that I have a fan curve enabled, and I have a good airflow case too with 8 case fans. Without the fan curve the max temp I’ve seen it go is only 73 degrees which is low for an fe even at 21C ambient(reviewers get around 75). My ambient is pretty high at 28-32C and thus my low temps for the gpu seem abnormal. I toggled the board power draw for the gpu in gpuz to show the highest reading and it was 325w. I’m pulling my hair out as to what’s wrong. Any input is appreciated. Thank you!

System:
Rtx 3080fe
5600x
Msi b550 tomahawk
Gskill ripjaws 3600mhz cl16 16gb
NZXT c750(psu)
Crucial p1 500gb ssd
 
Looks about right for a 3080. Check whether the card is running on 16x pci. Also pci gen4 might make a difference for x570 boards.
Also check if the memory is overheating, that’s a thing with the newer cards.

overall the score is close enough to fall under silicon lottery, might just be unlucky.
 
You have to remember there's a bunch of settings you can set with nvidia control panel for benchmarks.
Go to Nvidia Control Panel by right clicking your desktop, go to Manage 3d Settings.
Max Frame Rate Off
Power management mode set to Prefer Maximum Performance
Texture filtering - Quality High Performance
then click apply
See if that ups your score abit.
 
Spiritwolf said:
Looks about right for a 3080. Check whether the card is running on 16x pci. Also pci gen4 might make a difference for x570 boards.
Also check if the memory is overheating, that’s a thing with the newer cards.

overall the score is close enough to fall under silicon lottery, might just be unlucky.
Click to expand...
Hello, b550 has gen4 too. Going to the gpu and one ssd. And how do I check if the card is at 16x? And I can’t find ram temps in Hwinfo either. When you say closer to silicon lottery, meaning that it’s close to being a great chip and rn it’s just average? Thanks for the reply!
 
Hakanese said:
You have to remember there's a bunch of settings you can set with nvidia control panel for benchmarks.
Go to Nvidia Control Panel by right clicking your desktop, go to Manage 3d Settings.
Max Frame Rate Off
Power management mode set to Prefer Maximum Performance
Texture filtering - Quality High Performance
then click apply
See if that ups your score abit.
Click to expand...
Max frame rate is off by default and vsync I turned it off. Setting it to max performance will definitely boost my scores but I doubt the other people at stock did this. I just wanted it to pass out of the box tests without much tweaking around.
 
That’s a normal score for stock.

Your cpu may be holding you back a little bit but my 3080 went from 17k to 18.5k by undervolting.

I’m also running 3x8 pcie slot so I’d expect another 5% 0.5k when I sort that aspect out.
 
ikinson said:
Aibs get even more. Fe was reviewed to get 17500-17800. I checked youtube and Reddit for this. Hope I’m wrong.
Click to expand...

Yes AIBs do get more, because they are tuned better out of the box and tend to get better thermals. You'll need to overclock your FE, however if your getting FPS your really happy with, forget about Timespy - it's not worth the hassle unless you are aiming for the top x% and are getting paid for it. Use the benchmark to make sure you're about right, which you are :) if it was 13-14k then I'd worry.

Enjoy your card and the games :)
 
Brazo said:
That’s a normal score for stock.

Your cpu may be holding you back a little bit but my 3080 went from 17k to 18.5k by undervolting.

I’m also running 3x8 pcie slot so I’d expect another 5% 0.5k when I sort that aspect out.
Click to expand...
How does a 5600x hold back a 3080? Also, you got 17,000 at stock with 3x8? Mine is at the full 16x. I just checked in gpuz. Maybe my cpu is the problem as you said, don’t know if that cpu score is bad or good.
 
ikinson said:
How does a 5600x hold back a 3080? Also, you got 17,000 at stock with 3x8? Mine is at the full 16x. I just checked in gpuz. Maybe my cpu is the problem as you said, don’t know if that cpu score is bad or good.
Click to expand...
Well this is a 1440p test so higher fps can win out and it’s not reliant on single core speed. Simply put a higher cpu score does seem to improve gpu scores slightly.

My old 2700x used to give me a 10k cpu score, when I upgraded to a 3800x this improved both my cpu score and consequently my gpu score.
 
Bearing in mind that the 3080 is always limited by the CPU and the rest of the system at 1440p. Forgive me for sounding patronising, but it's amazing the number of similar posts I am seeing, that the 3080 is not performing, when it is not really being used in it's best environment, that is driving a high resolution monitor (higher than 1440p).
 
Brazo said:
Well this is a 1440p test so higher fps can win out and it’s not reliant on single core speed. Simply put a higher cpu score does seem to improve gpu scores slightly.

My old 2700x used to give me a 10k cpu score, when I upgraded to a 3800x this improved both my cpu score and consequently my gpu score.
Click to expand...
10k cpu score for the 2700x? My 5600x gets a 8286 lol this normal? and yeah you’re right. But can a 3700x outperform a 5600x in timespy just because it has more cores? Sorry, new to timespy. I usually use heaven only.
 
montymint said:
Yes AIBs do get more, because they are tuned better out of the box and tend to get better thermals. You'll need to overclock your FE, however if your getting FPS your really happy with, forget about Timespy - it's not worth the hassle unless you are aiming for the top x% and are getting paid for it. Use the benchmark to make sure you're about right, which you are :) if it was 13-14k then I'd worry.

Enjoy your card and the games :)
Click to expand...
Sorry I couldn’t reply earlier, I’m a newbie so only allowed to post 15times a day. So anyways, yes I guess I’m just nitpicking. I’m really happy with the performance in games. Got around 190-200fps in Witcher 3 1440p!
 
varkanoid said:
Its normal. Mine was the same till I tweaked and undervolted it.

Graphics Score 17 425

Dont always take what reviewers say as gospel.

With a bit of tweaking I now get 18118 on my last run.
Click to expand...
I see, as for undervolting, does it yield better scores only in synthetic benchmarks? Or does it give more fps too?
 
ikinson said:
I see, as for undervolting, does it yield better scores only in synthetic benchmarks? Or does it give more fps too?
Click to expand...

Depends on the card, if it is throttling speed due to hitting power limit then lower volts may allow it to reach higher clocks before hitting power limit and thus more performance. Lower temps from lower volts can also help a little.
 
ikinson said:
10k cpu score for the 2700x? My 5600x gets a 8286 lol this normal? and yeah you’re right. But can a 3700x outperform a 5600x in timespy just because it has more cores? Sorry, new to timespy. I usually use heaven only.
Click to expand...

That's right, the CPU test is heavily threaded so more cores (even slower ones) will result in s higher score.
 
You must log in or register to reply here.
Back
Top Bottom