• 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.

[SOLVED] MSI RTX 3080 low-ish Time Spy score in comparison to others?

I remember going through a similar scenario with my GTX 1080 my first 12 months of ownership were spent tweaking and swearing to god that mine is slower than everybody elses.

But after a while I realised all these reviewers had either crazy overclocks, always ran maximum components at the time, ran it on an open test bench etc etc.

Really ruined my first part of my experience with the card as I always felt I was behind. Turned out I really wasn't and my card was performing to spec and was decent on a manual oc too. Placebo can work opposite too in convincing you something is wrong no matter where you look or what you do. Once you go down this hole it's hard to drag yourself back out to where things seem right, I sympahise.
 
I remember going through a similar scenario with my GTX 1080 my first 12 months of ownership were spent tweaking and swearing to god that mine is slower than everybody elses.

But after a while I realised all these reviewers had either crazy overclocks, always ran maximum components at the time, ran it on an open test bench etc etc.

Really ruined my first part of my experience with the card as I always felt I was behind. Turned out I really wasn't and my card was performing to spec and was decent on a manual oc too. Placebo can work opposite too in convincing you something is wrong no matter where you look or what you do. Once you go down this hole it's hard to drag yourself back out to where things seem right, I sympahise.

Thanks dude, that actually helped a lot :)
 
Finally can post again.

After many suggestions and different steps, nothing has really worked with stock settings at 1440p. I have now managed to get a score of 17,018 in Time Spy but that's OC'n the GPU with: Core clock +110, Memory +400 and Power Limit 102%. This sustained an average clock of 2008mhz which is probably why it got just over 17k. Temps reached a max of 70c which is not too bad. Didn't have any crashes with this OC either. Not sure what the deal is in the end with my card since I still think stock out of the box it should perform better but there's too many variables to why it isn't performing. Port Royal score on the other hand was 11,585 Stock settings, that actually seems pretty decent so maybe my CPU is bottlenecking me somehow. For ref my CPU is 9700k running at 4.8ghz on all cores, any suggestions for the CPU are more than welcome :). I've already tried default settings in the bios, didn't make any difference unfortunately.
 
What temperature & clock does it run at ? I know another 3080 MSI owner having issues due to their card thermal throttling almost immediately
 
What temperature & clock does it run at ? I know another 3080 MSI owner having issues due to their card thermal throttling almost immediately

73c max if I leave it on its auto fan setting and with a custom fan curve 62c. Have tried with both options and the score didn't budge. Even with the +110 on the core clock temps only went to 75c
 
+400 on the memory might be making the score worse.
If the memory OC is too high then error correction kicks in which makes it slower. It won't crash or artefact so you won't know
 
+400 on the memory might be making the score worse.
If the memory OC is too high then error correction kicks in which makes it slower. It won't crash or artefact so you won't know

I thought that too but I tried it with +400 and with +100 and without and only with +400 would I get 17k score
 
8 full cores at 4.8GHz shouldn't be a bottleneck.

This is why it's confusing as the CPU score is around 8,200/above which I believe is correct for the 9700k. What's more confusing is Port Royal is chucking out numbers such as 11,585 on stock settings which as far as I'm aware is actually good and about right for the 3080.
 
Rather that killing yourself trying to reach synthetic benchmarks where an OC can give you 500 points, but in games might equate to 3fps, check how your doing on some games if you can.
 
Rather that killing yourself trying to reach synthetic benchmarks where an OC can give you 500 points, but in games might equate to 3fps, check how your doing on some games if you can.

In games it seems to perform perfectly fine, Warzone I average around the 140-170 mark which seems to be about right, that's at 1440p everything maxed raytracing off. Shadow of the tomb raider on the other hand is the opposite, it under performs terribly in comparison to the reviews. It's not so much that I'm having a terrible time in games because tbh it's great, it's that I wanted to check (Which I instantly regret :P ) what my card was like in comparison to others since there aren't many around atm and that's when I noticed I'm getting low numbers, I really want to find the cause that's all.
 
In games it seems to perform perfectly fine, Warzone I average around the 140-170 mark which seems to be about right, that's at 1440p everything maxed raytracing off. Shadow of the tomb raider on the other hand is the opposite, it under performs terribly in comparison to the reviews. It's not so much that I'm having a terrible time in games because tbh it's great, it's that I wanted to check (Which I instantly regret :p ) what my card was like in comparison to others since there aren't many around atm and that's when I noticed I'm getting low numbers, I really want to find the cause that's all.


Well if I ever get mine :sob I'll see how I go compared to you. I'm sure we'll figure it out.
 
I've noticed that you average clock frequency on the score says 1,730 MHz, I think that might be causing the low numbers also, Just did another run and got 16,428 after updating my bios to check and average clock freq was 1914 MHz....


In Firestrike it averages 1879.
 
My first thought after reading your first post was to instantly remove Dragon Center!

Weirdly I just looked up Dragon Center and removed it about 5 mins ago, tested again no difference, still 16,386. Somethings gotta give with this. There was only 1 time I got over 17k stock clocks from a cold boot and hadn't uninstalled/changed anything from that point. It can't be temps since I've tested with 100% fan speed and side panel off
 
Wow, hold up. Ok so I made sure that 'Prefer maximum performance' was enabled in nvidia control panel and then I thought, I haven't changed anything windows side so changed the power option to one called 'Ultimate Performance' and low and behold:

GPU score - 17,975 at stock settings, only thing I have is my custom fan curve in Afterburner

https://www.3dmark.com/3dm/51307035?

Is there something suspicious going under hood after the MLCC vs SP caps fiasco, I know the newer driver was supposed to keep the clocks from ramping up all the time but thing is, I never had an issue with my card and crashing above 2ghz???

Edit: What's even more weird is my average clock freq for that run was 1899mhz, lowest I've seen before. Gonna run another test to check if it's around the same score.

Edit 2: Ran it again with nothing changed and it's now a score of 18,034

Edit 3: Changed back the nvidia control setting to 'Optimal power' and changed the fine tune settings in Windows 10: PCI Express Link State Power Management to Moderate power savings and Processor power management - Min state 5% and Max 100%. Got a score of 18,006, still stock settings. I have noticed however that the LEDs are off my MSI card, probably just cause I uninstalled Dragon Center but very weird coincidence. Gonna reinstall it and see what happens.

Edit 4: Reinstalled Dragon Center to enjoy the leds and it seems to have made no difference to the score, GPU score 18,131. I am so happy right now, will carry on testing to establish it's for real and not some weird thing going but maybe Windows 10 power management is the issue? Would be interested to know if this make a difference with other people and any card I guess.
 
Last edited:
I have now concluded my tests and indeed setting my power management back to balanced decreases the score to 16k, putting it back to Ultimate Performance but with some adjustments to stop it running at 100% all the time has now increased to post over 18k results. I'm so happy that I have found the issues, as simple as it was. Please refer to my latest edit in the original post for more info :)
 
Back
Top Bottom