I tried everything | TheaterFire

[-]

postsshortcomments@reddit

Have you checked your temps? A 9070XT has a really high throttle temp, but usually youre looking for under 90-95C on equipment.

If you did not start with a fresh windows install, uninstalling all old drivers is as important as installing new ones. This includes chipset drivers, old GPU drivers, etc., A Windows reinstall basically ensures this happens (and sometimes it's fully necessary).

Your monitor is plugged into the GPU and not motherboard I/O, right? Motherboard i/o = CPU integrated GPU (when your cpu has an iGPU). GPU itself = GPU itself

Are you on the most recently updated BIOs?

With your PSU PCIe cables, there are what are called "PCIe pigtails" which is basically a PCIe cable that splits off into two headers. In especially older or budget PSU models, sometimes the PCIe pigtail has a max load across the cable. Modern, quality PSUs often have enough wattage to each connector. If you have enough unique, separate PCIe cables, plug each cable into the GPU just once. (If it's something like an A or B tier 2024 or later model there's a good chance that both headers can be plugged in, but if you have enough unique cables it's a troubleshooting step I'd take).

Someone mentioned you may have been scammed by buying a "bad used GPU." A 9070 XT should still be under warranty to the original buyer. In Europe, warranties are typically tied to product, are transferrable, but you may require proof of purchase. If you still think the model is broken and it was a possible scam, you may with to try to contact whoever you bought them to and explain this to them without being accusatory. Message them and ask them if they can help you by either providing a proof of purchase to redeem the warranty or offer it back if they want to do it themselves. People are usually more inclined to provide you with an hour of their time for $20, so that may be a good way to open that dialogue.

[-]

Disastrous-Media6050@reddit (OP)

Thanks a lot for the detailed reply. I’ll try to be detailed aswell.

My gpu didn’t even go over 60 degrees once. What was quite weird was the gpu tickrate and the voltage. The gpu tickrate got to over 3000mhz and the voltage to about 1100mV. I believe this is too high for a game like valorant even with uncapped fps. Maybe some sort of leak? Or some sort of problem with riot vanguard?

I did start with fresh windows, as I built the pc yesterday. However I think my next step would be to reinstall windows again. My game worked for like a round, then I did some other troubleshooting and everything after that had the stutters. My hope is that some hub or app I installed made the issue here

To the stutters: they are incredibly consistent. The game runs perfectly fine for roughly 30-60 seconds, after which the micro stutters appear. My gpu seems to be underclocking itself when the stutters happen

My monitor is plugged in correctly

Updated bios aswell, including every single driver for everything

This one is also kinda interesting: with my be quite pure power 13 m 1000w there were only two usable VGA cables (https://ibb.co/nqr2YtL1). I connected one 6+2 pin per cable each. Is that correct? Don’t really think there is a better way to do it with what I got

I don’t believe I got scammed. I did pay a relatively low, but still fair price for the card and the guy transferred his warranty to me. I also have the receipt for it now. He even told me he’ll help me for free if I have any issues with the warranty here.

[-]

postsshortcomments@reddit

You may want to also check your CPU utilization. If it's a BIOs related issue, there's a good chance it's actually a CPU-bound issue resulting in low-GPU utilization. So make sure you BIOs update. It's very common for early revision motherboards to have issues with CPUs that result in massive framerate drops/stutters.

If you have an old blank old SSD you can "save" your Windows drive to troubleshoot. Just remove it and install the OS to another drive. You can probably get away with an HDD even in the right conditions - but Valorant wouldnt be my first choice to troubleshoot it. I'm not very familiar with Valorants inner workings, but some cash shop/skin titles often have "on-demand texture streaming." Given that it's build for competitive play, I'd kind of doubt it but in some titles like Fortnite skins/player models don't read or load into VRAM until you encounter it the first time. Because of that, an HDD could cause unpleasant experiences (once loaded, it should perform equally.. but some larger installs & especially open world titles regularly load assets behind the scenes during gameplay). If you can identify ~15-30GB not open-world 8GB VRAM-era single-player title like Risk of Rain that doesn't have crazy HDD demands, an HDD install could be feasible if you're a bit patient and have chores or other things to do. Still, I'd expect a tiny bit of stutter with a "small install," but something like Palia runs perfectly smooth on an HDD.

Damn… how can I verify whether that’s true or not?