(PROBLEM SOLVED) The system frequently powers down during heavy tasks yet remains stable under lighter conditions.
(PROBLEM SOLVED) The system frequently powers down during heavy tasks yet remains stable under lighter conditions.
I am not planning to label this answer yet, since the issues continue. I intend to purchase a more powerful power supply for the setup, hoping this resolves the matter. I’ll revise and post a reply once confirmed. Thanks. My gaming rig is decent but outdated. Around 25 minutes in, the system freezes, the screen turns completely black, and my audio produces static. When I restart using Furmark, it abruptly shuts down and restarts. During gameplay, crashes occur, forcing me to power off by pressing the power button or switch. Event Viewer note: I install hardware and software myself, everything is set to default, never modified. The overall system is 5 years old; my PSU is 4 years old, and the GPU is 2 years old. Originally it was: CPU Intel i7 8700K, cooler CORSAIR Hydro Series H115i, RAM Corsair 32GB DDR4, GPU MSI GTX1060 (now Zotac RTX3090 OC since 2021), PSU Corsair RM1000i, OS drive on m.2 500GB, 2TB HDD, and a 250GB SSD in the back. CASE: Corsair Crystal 570X with RGB fans (3x exhaust, top and rear, plus 2x 140mm in-cooler fans). Screen: AOC C32G2ZE/BK 240 at 240Hz. Added external gear: Focusrite Scartlett Solo 2nd gen. Temperatures were high—CPU reached 70°C before rework; after adjustments, stayed around 65°C under load, 35–40°C idle. GPU heats up quickly under stress; with Firestorm’s default fan curve it hits 70–73°C, dropping to 60–62°C with a proper curve. I’ve tried several fixes: GPU repasted (possibly unnecessary), CPU repasted, fans repositioned, and even used a thermal paste. The PSU and motherboard are still functional. I’ve run SFC scans via command prompt, possibly due to software issues. Drivers and updates are current. Nvidia released a recent update (12 Dec) but hasn’t installed it yet. Anyone know how to resolve this? I’m puzzled—can type for hours without crashing, yet games fail after 25 minutes. I’m considering swapping fans for better cooling (Noctua or similar) and possibly upgrading the case with more airflow. There’s also a possibility of triggering an OCP (Over Current Protection), which is why I’ve been limiting GPU power via Furmark. **UPDATE** This was posted on 12th December on Reddit. Today I attempted to cap GPU usage at 95% using Zotac Firestorm, but the stress test ran over 25 minutes without shutdown. Now I’m running a new Furmark stress test with afterburn as a power limiter. If I don’t reply soon, I might be too tired to respond. I’m from Belgium.**
Hi, yes, when I assembled this PC five years ago I ignored the orientation, that was a beginner error. I have five fans—two at the front push fresh air in, and the top and back have exhaust fans, while the bottom has none. I plan to add one later. Back then the system worked perfectly without any issues. If there are any airflow problems, they should be resolved. Everything is checked and positioned correctly.
This seems to point to a power supply concern. The fact that limiting CPU power resolves the issue suggests the problem lies upstream. I've noticed failing PSUs and motherboards in similar cases. Given your test results, it makes sense the issue is likely with the power delivery chain—PSU, motherboard, or CPU. Since both components are reputable, it's hard to pinpoint exactly. Having test parts from someone else would help confirm. Ideally, I'd focus on the motherboard first, as CPU load usually doesn't exceed the full capacity of a 1kW PSU unit. This leads me to suspect the board's VRMs are under stress.
Hi, the CPU isn’t being limited, only the GPU is affected. I ran a Furmark for 37 minutes and it worked fine. That points to a power problem, though a 1k power supply seems excessive for a 3090 OC. I should have chosen a better one when buying the PSU. The machine shifted from heavy to light use based on my workload. There’s an active fan curve; it kicks up to about 70% around 55°C, but temps stay around 60°C. I’ve used the 3090 for two years without issues or overheating. When launching games like Construction Simulator, it runs smoothly for about ten minutes before shutting down completely, turning black. But with Furmark running the GPU at full capacity for 37 minutes, it doesn’t crash—just stops abruptly.
It seems you were trying to discuss GPU power requirements and PSU performance. The issue might stem from an insufficient power supply unit. I remember a 3090 that struggled under high loads, crashing during intense gaming or synthetic tasks. It wasn't about the total output but how quickly it could manage sudden power demands. Upgrading from a 1kW to a similar model resolved the problem. Your situation might differ, so testing with a loaner PSU could help confirm.
I'm planning to explore options for a fresh power supply. Sadly, there aren't many local shops or people who lend PSUs around here. Likely I'll have to purchase a brand new one. My budget is around 1200 to 1600 watts, and I'm considering upgrades later with better components.
It’s helpful to use the bigger PSU, though newer ATX3.0 models seem better at managing spikes. I’m currently running a 1kw SFF ASUS PSU that powers a 14900k and an RTX4090 without any stability problems. No overclocking needed.
The 14900k works well. It’s basically a marginally more efficient 13900k, which is why I chose it. I reduced my system size so I sold the 13900k to a friend and kept the 14900k for a lighter load on my smaller cooler. If I’d started from zero, I’d go with a 7800X3D setup (which also fits SFF nicely). However, I wouldn’t build a brand new Intel board right now since there’s no upgrade route. If you’re starting fresh, opt for AM5; otherwise, hold off until Intel releases a new socket.