Question New gpu/cpu far below average on 3Dmark tests. Advice appreciated SOLVED! :D

[EannaS]

Hi all,

Been helped a few times on this site and I’m really hoping this can be another. I’m at my wits end really, I’ve just recently upgraded my pc, everything except the motherboard almost, and my timespy and Firestrike scores are well below average for my 4070 super ti and 5800x3D. I’ve uninstalled previous drivers and reinstalled the latest, enabled xmp in bios, newest bios version etc.. I genuinely can’t find anymore posts or information on this specific issue so I hope someone can help. I did try undervolt both cpu/gpu a small bit but I’m unsure if I even did it correctly honestly, I’m overwhelmed. I just want to get my moneys worth, and the loudness/sound of the pc is another big issue for me. I’ve set fan curves etc, but issues persist. I’m debating refunding everything and trying to rebuild my old setup. (2060, 5600x) but if I can get this working obviously I’ll stay with this new pc.

If anyone has any advice on getting my gpu/cpu to stock performance and an average score on timespy I’d be very happy. For reference my timespy score was 17k while the average was almost 21k, and my firestike was 34k with the average being 40. If someone could work with me gradually in getting the pc workable for my specific needs and sensitivities also I would be in your debt.

My specs:
Case

COOLERMASTER MASTERCASE H500
ARGB GAMING CASE

Processor (CPU)
Ryzen 7 5800x3D

Motherboard

ASUS® TUF GAMING B550-PLUS WiFi
(DDR4, USB 3.2, 6Gb/s) - ARGB Ready!


Memory (RAM)
32GB Corsair VENGEANCE DDR4
3200MHz (4 x 8GB)


Graphics Card
Zotac black trinity 4070 super ti 16GB vram

CORSAIR RM850x 80 Plus Gold PSU
Thank you in advance.



links to 3Dmark graph images -

https://ibb.co/Fmfr0dk

https://ibb.co/KLf65Ft

https://ibb.co/0tQDgQs

image hosted at ImgBB

Image image hosted in ImgBB

ibb.co

https://ibb.co/mH4QzfW

 

[--SID--]

Besides the CPU and GPU, what's the brand and model of the other parts includes the PSU?

 

[Nine Layer Nige]

Hello, more info would help...
Please list full pc specs for the community, eg.

CPU:
GPU:
Mobo:
Mobo BIOS Version:
Ram:
Ram configuration, slots, xmp (on/off) :
PSU:
Storage:
Windows version:
Monitor screen resolution refresh rate:

CPU / GPU temperature ?

 

[thestryker]

Definitely going to need more specifics on the breakdown of the scores as well. Is it one, the other, or both CPU and GPU underperforming in the tests.

 

[EannaS]

Besides the CPU and GPU, what's the brand and model of the other parts includes the PSU?

Sorry I’ve amended that now.

 

[EannaS]

Definitely going to need more specifics on the breakdown of the scores as well. Is it one, the other, or both CPU and GPU underperforming in the tests.

Hi, it’s more so the gpu I believe is underperforming. I’ve updated my original post with my specs. I also have the monitor graphs of the score test maybe it could help you see it better, if I manage to post them here somehow.

 

[EannaS]

Hello, more info would help...
Please list full pc specs for the community, eg.

CPU:
GPU:
Mobo:
Mobo BIOS Version:
Ram:
Ram configuration, slots, xmp (on/off) :
PSU:
Storage:
Windows version:
Monitor screen resolution refresh rate:

CPU / GPU temperature ?

Hi! Thanks for answering!

Sorry, I’m new here and not the most technologically minded. I put some of my specs in the og post. Hopefully it helps!

The bios version is the latest possible I could get, released in 2021. The ram, I turned xmp on, storage I have a 4TB HDD, and a 2TB HDD, a 2TB SSD, and a 500GB SSD boot drive.

Windows version is the newest windows 10. My monitor and current refresh rate is 144hertz.

The cpu/gpu temperatures seemed good but I’ll have some of the monitor graphs from the score test if that would help you? I’ve posted them in a different forum to no avail.

 

[Nine Layer Nige]

Temps seem fine.
You might try disconnecting any of your HDD storage if its non essential and rerun.
Also, try running only 2 sticks of ram in the primary slots instead of 4 and rerun tests.

 

[EannaS]

Temps seem fine.
You might try disconnecting any of your HDD storage if its non essential and rerun.
Also, try running only 2 sticks of ram in the primary slots instead of 4 and rerun tests.

I see, alright I’ll try that! Thanks! It’s not a long term fix hopefully, I do use both hdds for games but I might try to transfer some of them to the new ssd if it helps my gpu perform better. As for the ram, I only just got new ram a while ago so I probably won’t upgrade for a while. Hopefully it isn’t too drastic of a difference but if it helps narrow down my issue I’ll try this absolutely.

Should I have Mai afterburner running for the test? It’s slightly undervolted but I’m not sure if it would affect it or not? Thanks

 

[Nine Layer Nige]

Sorry, I cant advise about afterburner as I have never used it.

 

[EannaS]

Sorry, I cant advise about afterburner as I have never used it.

No problem, I’ll try those suggestions with and without afterburner and I’ll get back to you here thanks so much again, I really just want to make some progress with it and i appreciate any help possible.

 

[Dragos Manea]

Be sure to turn off g-sync during benchmarks, uninstall old video driver with DDU

 

[thestryker]

Frequencies and temperatures both look good so there's nothing obvious on that side of things. While the X3D parts aren't particularly memory sensitive it would be worth making sure that the memory is running 3200. You can use CPU-Z and check on the memory tab where it should show DRAM frequency which should be showing ~1600.

On the results page 3DMark should have the following, and both should be disabled as they're known for messing with performance:
VBS status: Disabled
HVCI status: Disabled

Ensure the BIOS is actually the latest which is:
TUF GAMING B550-PLUS (WI-FI) BIOS 3404
Version 3404
16.14 MB
2023/10/31

MSI Afterburner will have negligible impact on any testing scores and certainly nothing to the degree that seems to be going on here.

 

[EannaS]

Frequencies and temperatures both look good so there's nothing obvious on that side of things. While the X3D parts aren't particularly memory sensitive it would be worth making sure that the memory is running 3200. You can use CPU-Z and check on the memory tab where it should show DRAM frequency which should be showing ~1600.

On the results page 3DMark should have the following, and both should be disabled as they're known for messing with performance:
VBS status: Disabled
HVCI status: Disabled

Ensure the BIOS is actually the latest which is:
TUF GAMING B550-PLUS (WI-FI) BIOS 3404
Version 3404
16.14 MB
2023/10/31

MSI Afterburner will have negligible impact on any testing scores and certainly nothing to the degree that seems to be going on here.

I see, that’s great! Some people thought I’d have to return the card, but it sounds like the card is working well just low scores.. I hope together we can find out why! Yes my ram is running at designed speed. My bios is that version I believe, and I’ll try disabling those options in 3Dmark. Thanks so much, if you have any other suggestions I’d love to try them.

 

[EannaS]

Frequencies and temperatures both look good so there's nothing obvious on that side of things. While the X3D parts aren't particularly memory sensitive it would be worth making sure that the memory is running 3200. You can use CPU-Z and check on the memory tab where it should show DRAM frequency which should be showing ~1600.

On the results page 3DMark should have the following, and both should be disabled as they're known for messing with performance:
VBS status: Disabled
HVCI status: Disabled

Ensure the BIOS is actually the latest which is:
TUF GAMING B550-PLUS (WI-FI) BIOS 3404
Version 3404
16.14 MB
2023/10/31

MSI Afterburner will have negligible impact on any testing scores and certainly nothing to the degree that seems to be going on here.

Hi, I did all of what you said as well as undoing all my undervolts for everything and resetting the bios (and re-enabling xmp), I got a much more average/decent score 20361 when the average is 20831. I’m far happier with this, I know it’s working as intended and I don’t have a faulty or waste of money card! I just need to sort out my fan curves so they don’t annoy the hell out of me and maybe undervolt slightly and test if it ruins performance again. Thanks again for your help.

 

[Nine Layer Nige]

Great news

 

[EannaS]

Great news

Thanks for all your help!

I think what it was, was my undervolt was too aggressive, it lowered my score from 20k which is the average to only 17k. I’m just not using an UV now, until I need it or can justify the performance loss. Maybe I did it wrong or too much though.

 

[Avro Arrow]

Thanks for all your help!

I think what it was, was my undervolt was too aggressive, it lowered my score from 20k which is the average to only 17k. I’m just not using an UV now, until I need it or can justify the performance loss. Maybe I did it wrong or too much though.

Well, it depends... Sure, it might show a 15% loss in a synthetic benchmark but benchmark scores don't necessarily translate into gaming performance. The difference might be imperceptible when gaming and if that's the case, then the undervolt would be totally worth it.

 

[EannaS]

Well, it depends... Sure, it might show a 15% loss in a synthetic benchmark but benchmark scores don't necessarily translate into gaming performance. The difference might be imperceptible when gaming and if that's the case, then the undervolt would be totally worth it.

when researching undervolt I saw constantly that it wont affect performance at all, for a standard UV anyway. And I’m sure it’s doable, that’s just how I got my score back up to an average/normal % which I wanted so I can feel relieved my card is ok and working properly. Probably down the line I will experiment with undervolting it just to help temps and lower fan speed a bit, but for now I’m happy enough with it using how it was designed and figuring out how it works. Even on helldivers 2 it can’t really keep 144fps on 1080p, (it drops to the 90s sometimes, I may try cap it at 120 so it’s less noticeable), but I see a lot of ppl struggling with that game it seems. I know benchmarks don’t really mean a lot, they’re overly inflated, I only care that I got what I paid for and it helped in that sense, I probably won’t benchmark again unless I undervolt. Thanks for answering

 

[Avro Arrow]

when researching undervolt I saw constantly that it wont affect performance at all, for a standard UV anyway. And I’m sure it’s doable, that’s just how I got my score back up to an average/normal % which I wanted so I can feel relieved my card is ok and working properly. Probably down the line I will experiment with undervolting it just to help temps and lower fan speed a bit, but for now I’m happy enough with it using how it was designed and figuring out how it works. Even on helldivers 2 it can’t really keep 144fps on 1080p, (it drops to the 90s sometimes, I may try cap it at 120 so it’s less noticeable), but I see a lot of ppl struggling with that game it seems. I know benchmarks don’t really mean a lot, they’re overly inflated, I only care that I got what I paid for and it helped in that sense, I probably won’t benchmark again unless I undervolt. Thanks for answering

Sure, there's nothing wrong with running a card at stock. After all, that is how it was meant to be used and will give the most stability. I've been thinking of doing an undervolt to my RX 7900 XTX because it can be done through the Adrenalin Software Suite and if it makes my card run cooler, so much the better. It definitely has the performance to spare!

 

[EannaS]

Sure, there's nothing wrong with running a card at stock. After all, that is how it was meant to be used and will give the most stability. I've been thinking of doing an undervolt to my RX 7900 XTX because it can be done through the Adrenalin Software Suite and if it makes my card run cooler, so much the better. It definitely has the performance to spare!

Definitely go for it! As people say there’s no damage risk with undervolt as compared to overclocking. The worst thing is maybe crashing until you lessen the amount a bit through trial and error. Good luck with your card!

 

[Avro Arrow]

Definitely go for it! As people say there’s no damage risk with undervolt as compared to overclocking. The worst thing is maybe crashing until you lessen the amount a bit through trial and error. Good luck with your card!

Thank you! I wish the same to you.