Severe failure under intense energy use
Severe failure under intense energy use
Hello everyone. I'm having some hard crash when our new setup uses a lot of resources. I can consistently make it crash by running y-cruncher on the FFTv4 test. It maxes the CPU and RAM, then crashes about 5-30 seconds later. Some games like "Baldur's gate" or "No man sky" crashes after some playing too. Max temp around 60 deg on GPU and 50 on CPU. Max 650w of power consumption on heavy benchmark. We've been troubleshooting for about a months now and we still can't make it work like we want. We don't care that much about performance at this point, we just want to be able to run games without any crashes Here is the setup TUF GAMING X870-PLUS WIFI AMD Ryzen 7 9800X3D RX 7900 XT (Sapphire) Ripjaws M5 Neo RGB 2x32GB DDR5 EXPO+XMP (F5-6000J3636F32GX2-RM5NRK), placed on A2 and B2 Corsair Shift 1000w gold For my drivers, I used BIOS 1022 AMD Adrenalin for graphical card AMD Adrenalin for CPU Asus TUF Gaming drivers for chipset Asus TUF Gaming drivers for all other components included in the motherboard What I tried to fix this issue, without success Enabling/Disabling EXPO Also tried BIOS from 1022 to 0816 RAM speed from 6000 to 3200 (Also tried 2600) 1/2 of the two RAM sticks. Also tried swapping their order Had my components tested in a Tech repair shop. They said they were fine. Full performance mode on AMD Adrenalin Mode Eco on AMD Adrenalin (worked a bit, some games don't crash with it, including "No man sky". Baldur's still crashes.) Tried this video : https://youtu.be/kJgcadbSAPM Make sure not to daisy chain any PSU cable Removed CPU, GPU, RAM, PSU cables then installed them again. AI Tuner. Sometimes made the PC stuck on post with White QLed Disabling Integrated Graphics Tried another Power supply (Corsair RM 750 Gold) - 26 april Using another Graphic card (GTX 1660) - 24 april Tuned to max 90% of CPU usage - 29 april I noticed that the CPU seems to clock at 5.2Ghz instead of staying to 4.7, even though there are no Overclocking enabled. Also, I don't have any other component to swap to try if they are the problems, expect for GPUs. Is there some thing I missed or that you think I should try? Thanks in advance! Edited April 29, 2025 by Alguewator Added the things I also tried on 29 APRIL
Hi! Thanks for the update. I also used a different PSU and the problem didn’t appear. I got the same crash with a Corsair RM 750 Gold (using integrated graphics) and also with an RTX 1660, but the issue remains.
We share quite a bit of the same configuration except I'm using a 7900 XTX. Both have the 9300X3D CPU on a 650 chipset board. I haven't encountered any of the problems you mentioned. I'm using AMD Expo profile, overclocked both CPU and GPU, and the only thing that stands out is the Expo profile, sine wave UPS, and the Asus gold 1000W PSU. I still notice a lot of GPU coil whine.
Thanks for the response! I'll attempt to boost performance with EXPO when I return to work. It seems these CPUs might be more reliable at higher speeds...
Thanks again for the response. I attempted to use the AI optimiser Overclock to check for improvements, but unfortunately nothing changed. I also tried Undervolting through the ADM Overclocking settings, yet still faced difficulties. FORTUNATE I discovered some hidden options that significantly improved stability and allowed me to pass a FFTv4 test without any crashes. Thanks to this Reddit post: https://www.reddit.com/r/GamingLaptops/c..._clock_to/. I also adjusted the power-saving settings on Windows, setting the maximum CPU usage to 95%, which capped the CPU at around 4.50Ghz. This seems to resolve the problem, though it will reduce performance. I’m not sure how noticeable the gain will be in gameplay, but I’ll keep an eye on it. For now, I don’t think this issue is fully fixed, but if it remains stable for a week, I’ll treat it as a solution.
Hello! Thanks for the guidance. After checking the bios, the feature was turned off by default. I attempted to turn it on but couldn't succeed. I also examined my RAM manufacturer's specifications and settings, but everything seemed correct. Following a Reddit thread, I managed to enable "EXPO + on the fly" by activating "Memory Context Restore." After trying it again, the test stopped with an error on Core 7 instead of crashing the whole system. That's a positive outcome since the PC won't fail completely. I plan to run it in-game to check for any further issues. Good luck!