Rog 4090 strix stuttering in all games

[Kayaba]

I bought asus strix 4090 graphics card on march 2024 it worked good at that time but after 3months iam getting stuttering in every game when the card was new my temps were core 69c, memory 76c and hotspot below 83c at full 500watt load but now at same 500watt load card is reaching core 77c, memory 78c and hotspot 100c and in hwinfo it shows thermal limit 84c and i too saw when the gpu cross 84c hotspot the gpu core clock reduces from 2760mhz to 2755mhz.

I tried everything changed psu, Ram reinstalled fresh windows and tried windows 11, windows 10 my cpu is 13600k it is not the one causing stuttering because i have 6800xt gpu I tried with that gpu no stuttering in any game that i play but with 4090 getting so much stuttering also tried old nvida drives and latest drivers but still getting stuttering don’t know what to do should i rma or what.

”I used DDU and selected clean install when swapping gpu to test stuttering dont tell that amd drivers causing it i removed amd drives with Display Driver Uninstaller DDU”

[Kayaba]

I bought the 4090 Gpu on march during summer ambient temp about 34c at that time my hotspot never went above 84c on 500watt load 

But now it is rainy season my room ambient temp 26C at the same 500watt load test hotspot hits 101C is that my problem i don't care about temps as it is under spec my main concern is that it is stuttering check the above frame time graphs

[Silent_Scone]

The thermals may be inducing the problem, but one can only speculate. Have you checked to see what's happening at the time of the latency spike? The easiest tool for this is GPU-Z. Check the Performance cap reason.

Have you considered repasting the GPU? In the climate you're in during summer, it's not unreasonable to assume the thermal application will dry out faster.

[Kayaba]

The stutter causes even when the hotspot temps are below 84c

The 4090 i have shows 24142MB vram but other 4090 have 25156MB can you share the screenshot by searching dxdiag  in windows search and goto display tab it shows the gpu display vram 

I didn't check the gpuz data while stuttering but i ran 20loops of 3dmark speed way benchmark here is the screenshot

[SeaSickLED]

I had the same problem. Forget all of BIOS etc. tweaking, the problem lies somewhere else. You will be able to reproduce the problem by simple usage of FurMark. Just Install FurMark and let it run on highest settings. And watch the hotspot temperature. Surest way to reproduce the crashes. My card reached 96°C very soon and - crashed. I ignored the warranty, opened the card (it's very simple) and observed only some few thermal paste on the GPU. Remove the old thermal paste, put triple amount of Thermal Grizzly Kryonaut Paste and watch your GPU get only 70°C hot. And you can happily restore your old BIOS Settings.

Have fun and good luck.