Is my 5820k Overclock Stable
Is my 5820k Overclock Stable
CmdrJeffSinclair responds with a similar explanation, noting that games don't seem to gain from higher speeds. He adjusted to 4GHz Auto-OC via BIOS and reset voltage settings. The 4GHz setting came bundled with the BIOS. He plans to run tests to confirm improvements. Voltage is back at 1.18v for Vcore, which he believes improves thermal efficiency significantly, potentially eliminating overheating issues even during extended AIDA64 tests.
You're experiencing stability and safety issues. This could mean the system crashes or encounters errors. Stable means it won't crash or throw exceptions. Safe means it won't damage my processor. For now, your processor seems stable. A longer test is needed to confirm. Prime95 can help assess CPU performance. Currently, your temperatures are elevated. During the recommended 8-hour test, keep temperatures below 86°C. Exceeding this could indicate unsafe overclocking.
In brief: no. These test results are intriguing but don’t reflect typical temperatures experienced by most users during regular use. As long as your temperatures stay stable and don’t cause throttling, you’re safe.
Despite this, I tend to be cautious and run my 4790k at a lower voltage. Two reasons: (1) I don’t require a significant boost in clock speed if performance demands increase, then it’s time for new hardware; (2) I want the setup to last me about ten years, and I’m very budget-conscious.
A slightly higher temperature on the two cores could suggest the thermal interface material is uneven. Core temperatures naturally vary, but anything above a few degrees is worth checking.
What cooling solutions are you using? What type of case are you employing?
Those clocks and voltages seem likely to stay stable... probably. That's the expected behavior at those clock settings. You might even manage to achieve a slightly lower voltage there. However, the best way to confirm is to run AIDA64 or OCCT for a minimum of 4 hours, preferably 6 to 8. If it works, it will be fully stable.
Those temperatures are higher than I'd like my system to handle, by the way... I wouldn't expose a processor to that daily.
Regarding Linus... he sounds intelligent, but he's clearly not very careful.
In summary: no. The torture test numbers are intriguing but don’t reflect typical temperatures experienced by most users during regular operation. As long as your temperatures remain stable without causing throttling, you’re safe. Still, I’m a bit cautious about this and run my 4790k at a lower voltage. Two reasons: (1) I don’t require a significant boost in clock speed if performance demands were higher, then it’s time for new hardware; (2) I want this setup to last me around ten years or more, and I’m very budget-conscious. A slightly higher temperature on the two cores could suggest the thermal interface material is uneven under the cover. Core temperatures can vary, but a few degrees above normal isn’t a major concern. 10 degrees or more would be worth checking. What cooling solutions are you using?
Those clocks and voltages seem likely to stay stable... probably. That's accurate regarding their expected behavior. You might even manage to achieve a slightly lower voltage at those clocks. However, the best way to confirm stability is to run AIDA64 or OCCT for a minimum of 4 hours, preferably 6 to 8. If it succeeds, it will be fully reliable.
The temperatures are higher than I'd prefer for my system, by the way... I wouldn't expose a processor to that daily.
As for Linus... he sounds intelligent, but he's quite naive. He's not the most knowledgeable person, yet he does a decent job explaining things to newcomers. Nowadays, though, he often comes across as more of a marketing figure than a genuine expert. He's sold out his ideas.
He spreads bad news that his followers accept as truth, since it comes from the magic box. It's similar to watching fake news or CNN and believing everything a "news reporter" says. Ish... nobody does research anymore—they're not curious about learning, just want someone to give them info that matches their existing ideas.
In brief: no. The temperature test results are intriguing, yet they don’t reflect the typical temperatures experienced by most users during regular use. As long as your temperatures remain stable and don’t cause throttling, you’re in a safe position.
Despite this, I tend to be cautious here. Running my 4790k at a lower voltage makes sense for two reasons: (1) I don’t need a significant boost beyond the chip’s capability, and if performance demands it, it’s time to upgrade hardware; (2) I want this setup to last me several years, and I’m budget-conscious.
The slightly higher temperatures on the two cores could suggest the thermal interface material is not perfectly uniform under the cover. Core temperatures can vary, but a few degrees is normal. If it rises by 10 degrees or more, it might be worth checking further.
What are your cooling solutions? What case do you use? Also, what type of cooler are you considering?
I’m running an i7 5820K @4.3GHz with ASRock Extreme4, CoolerMaster Seidon 240M, EVGA SuperNOVA G2 1000W, and Phantek Enthoo PRO.
I have ample airflow in the case—200mm intake fan at the front, a 140mm exhaust fan at the back near the CPU, and two 120mm fans mounted high up. The case stays cool, quiet, and dust-free. The top, front, and sides of the case are cooler than my bedroom. Only during stress tests does the top airflow near the cooler get warm, similar to bath water temperatures—not hot.
All fans are set to Custom mode, activating at full speed if temperatures reach 69°C. So far, no test has pushed speeds beyond 55% (set for 59°C and lower), then 1900rpm at 65% for 60°C–64°C, and finally 2400rpm at 69°C. Anything above that could risk the system reaching around 70°C. Since I build this PC myself and it’s close to my bed, I’m very aware of its behavior. I hear the fans running constantly and sense the PC’s condition intuitively. I’ve been building computers for nine years, but this is my first fully custom build. Others were simpler, mid-range models with basic Hyper-5 EVOs and no overclocking, so I’m not entirely confident in every aspect...
==PC behavior n a nuthsell==
At 4.3GHz, AIDA64 quickly reaches 80°C within about 7 minutes across all cores, and by 15 minutes, cores hit 83°C but stay there even after 20 minutes. The temperature doesn’t climb much longer, though it takes longer to reach 83°C at a 4GHz clock. At 4GHz it still stays around 81°C instead of 83°C after 20 minutes—just a small difference.
At 1.285v, the voltage is stable with no issues. I could try 1.275v if needed.
Games never exceed 61°C, remain stable, and don’t cause crashes or visual glitches.
Passed 36 hours of Memtest86+ with this overclock after a week of RAM timing tweaks; fans ran at 50% during testing, then returned to default since performance gains were noticeable.
3DMark11 is the only application that forces fans to spin at 80% RPM, but temperatures stay below 69°C. At 70°C, fans reach 100%, which is loud. My custom fan profile works well, letting me hear temperature changes clearly.
The cooler air stays cooler than my room, except during AIDA64 when the top vents warm up (not hot).
Any suggestions? Should I stress-test for another four hours to ensure temperatures stay under 80°C, especially if cores hit 80°C at full speed? I could tolerate a chip that’s not perfect if the cooler is solid and I’m comfortable with the noise.
Those clocks and voltages seem likely to be stable... probably. That's the expected behavior at those settings. You might even achieve a slightly lower voltage there. However, the best way to confirm is to run AIDA64 or OCCT for at least four hours, preferably six to eight. If it works, it will be fully stable.
The temperatures are higher than I'd prefer for my system, by the way... I wouldn't expose a processor to that daily.
As for Linus, he sounds intelligent but is clearly not very careful...
Those clocks and voltages seem likely to be stable... probably. That's accurate regarding the expected values at those clocks. You might even achieve a slightly lower voltage there. However, the best way to confirm is to run AIDA64 or OCCT for at least four hours, preferably six to eight. If it succeeds, it will be 100% stable.
The temperatures are higher than I'd prefer for my system, by the way... I wouldn't expose a processor to that daily.
As for Linus... he sounds intelligent, but he's clearly not very careful....