Does PC stop working under stress?
Does PC stop working under stress?
Recently, I acquired a budget i7 processor for my computer to minimize bottlenecks. My system would shut down because the CPU hit 95% or higher during stress tests, so I purchased an affordable cooler which now keeps temperatures below 70°C. When I disable turbo boost (keeping it at 3.4GHz) and run Prime95, the CPU uses between 52-53 watts while running at 3.4GHz and remains stable. It also supports normal gaming and work tasks with turbo enabled. Temperatures stay very low, under 70°C.
However, activating turbo boost (bringing it to 3.67GHz) causes the CPU to draw up to 66W during a Prime95 test, leading to a shutdown within about 30 seconds after starting the test. Video rendering at 4K also triggers a shutdown after roughly five minutes, with a peak power draw of 57W.
Despite this, normal stress tests show the GPU drawing less than 55W even under heavy load (110W max), suggesting the power supply unit is sufficient. Cooling for both CPU and GPU appears adequate. The case airflow is decent, and the GPU only reaches around 65°C with my custom fan curve.
I suspect the issue might lie with the motherboard, particularly the VRMs. It has a 4-pin CPU power connector, but only two chokes are visible on the left side, with one large choke on top. This is similar to what I experienced with my old i5, which also had limited turbo boost performance (max around 33W). My previous i5 didn’t have problems with turbo boost, but it wasn’t very efficient either. There are no motherboard temperature sensors, and HWINFO doesn’t display VRM temperatures.
Specs:
- Motherboard: Esonic H61 FEL-U
- CPU: i7 3770
- Cooler: Coolermaster Hyper T20
- RAM: 16GB DDR3 1600MHz
- GPU: Zotac Gaming X GTX 1660S
- Power Supply: Corsair VS650 650W 80+ Bronze
- Hard Drive: WD BLUE 7200RPM 500GB 16M
- SSD: 2 generic SATA SSDs 256GB (dram-free)
What i5 were you using before the i7-3770? Did you have to update the BIOS before installing the i7-3770? How long is your PSU? You didn’t mention the case model.
The board isn’t suited for high-end setups, it’s not built to handle top processors. If your BIOS needs an update, do it now. Reset the CMOS and try again with a stress test on your system.
Additionally, you might need another cooler. The T-20 is likely a compact tower model. Consider evaluating a good down-firing air cooler that not only cools the CPU but also directs airflow to the board's voltage regulators.
It seems the VRM might be the root cause, either due to lack of cooling or excessive heat. Consider using affordable heatsinks and/or fans to improve airflow. However, this might not fully solve the issue.
It was an i5-3470S (65w) with no BIOS update. It worked straight out of the box, and the BIOS seems quite basic with limited flashing capabilities—hard to believe a Chinese generic OEM would skip updates. The power supply is outdated, around three years old; I think it’s still fine and can handle full gaming loads without issues. The case is Antec NX200, featuring a 120mm exhaust fan. I installed a standard cooler fan on top as an intake, directing air straight onto the VRMs.
I'm having trouble getting a new cooler, but this one is working fine. I'm using a stock intel fan to blow air on the VRMs and also as an intake. I accidentally touched the VRMs with my finger and it got hot—probably around 50-80 degrees. I don't have a heat gun to check the temperature.
I'm experiencing a lot of noise from a stock intel fan, and it seems like it's not making much difference. Heatsinks are a good idea, but the cooler's circular mount is causing some interference.