Prime 95 performance issues
Prime 95 performance issues
Looks like this OC might be a bit shaky at first. Delidding could harm the CPU if done improperly and you overlook certain components—especially if you switch to liquid metal as a TIM. But if your CPU has held up for some time, it’s probably not broken.
At 6700K, the CPU runs at 4GHz with a maximum boost up to 4.2GHz. A 4.6 performance at 1.3v is a known improvement, but you should be aware that your custom settings might affect stability. Memory seems okay unless there’s a deeper issue; crashes in Windows usually point to the OS or memory tests, not the CPU itself. Thread problems are likely coming from your CPU’s handling of intensive tasks like FFTs, not the blend workload.
Small FFTs tend to draw more power for heat management, so if you push heavy AVX workloads, keep an eye on thermal limits. Blend is designed to stress all CPU and RAM components, including power delivery and IMC. It’s useful for identifying weaknesses in Prime95 tests.
ASUS RealBench is another solid choice, often used for stability checks across many setups. It simulates real-world scenarios and can expose issues you might miss otherwise. Sometimes stability only appears under sustained loads—try running a long test (hours) and see how it holds up.
If you notice crashes during Windows or general use, revisit your OC parameters and conduct thorough stability tests. After adjustments, let the software run overnight and observe if it remains stable. Many OCs struggle with voltage expectations and don’t perform consistent long tests, which can lead to unexpected failures.
Yes, most small FFTs tend to trigger errors sooner than Blend mode when the system is unstable. I reduced the clock speed by 100Mhz and am checking for stability now. Recently, I played Outer Worlds, which is a demanding game for hardware but performed well without issues. I spent many hours without any problems and most other games also ran smoothly. However, one UE4 title clearly showed instability. I plan to test its stability, but if downclocking doesn’t help, it might be due to the VRM MOS temperature reaching its maximum. Also, the machine stayed stable for a long time. The delidding was done poorly—no polishing was applied to protect the CPU components under IHS. I used only liquid metal and a loose lid. Still, if it were a short circuit, the issue would appear immediately in my view.
If it doesn't assist, test with stock clock speeds and voltages. If OC stability fails at those speeds, try again.
It works well for a brief stability check, spotting obvious problems fast. It’s not as consistent as Blend, but it’s faster when you’re setting up a new overclock and then fine-tuning once you reach your core limits. Yes. I can run Battlefield 4 and Battlefront 2015 on my X58 rig with an X5670 at 4.3Ghz, paired with SLI GTX 780s running between 1300-1320Mhz or at default speeds. In Destiny 2 it sometimes crashes unexpectedly (I had to boost CPU voltages, and I’m still doing a stability test overnight). You can achieve smooth performance across all games, yet still encounter random failures for specific tasks or workloads. That’s the reality of overclocking—what you expect isn’t always what happens in practice. Initially you’re let down, then you weigh your choices: lower clock speeds or increase voltages. Raising voltages raises temperatures; Intel’s safe limit for a 6700K is about 1.35V (if you’re cautious), while most 14nm chips can handle up to 1.4V (if you’re willing to try). Going lower cuts performance, but a jump of 100 to 200Mhz usually goes unnoticed in games. It looks like the VRMs are solid and you have large aluminum heatsinks, which usually means they can handle high heat. Unless your case lacks airflow, pushing the board to overheat the VRM is likely necessary. Reviewers have successfully run 7700K at 5Ghz on this board, but a 6700K at that setting is risky.
Initial checks suggest a few cores struggled with OC at 4.6Ghz, but stability returns around 4.5Ghz. I'll sleep through the night. Discovered that running CPU and GPU at full capacity with Prime95 and GPUPI causes temperatures to spike to 86°C. Poor ventilation is the issue. I didn't notice this before since I'd used 4.6Ghz for over a year without any problems. Thanks to @Zando Bob and @Vishera!
The cooler appears acceptable, but this particular situation represents the most severe airflow issue.