i5 6600k OC
i5 6600k OC
Question; what's the most effective method to check stability?
I'm hesitant about keeping the computer running continuously for 24 hours with Prime95, so... I'm looking for something faster.
I've used Aida64 (1-2 hours), Intel Burn Test (50 runs at high speed), cinebench, realbench.
All of them passed without issues. Temperatures varied across each program... Aida64 gave the highest readings, the Intel Burn Test was the poorest, reaching 90°C on one core!
All tests were done at 4.8GHz with 1.36V (stock voltage)...
Today I was watching football on one screen while playing a game on another - a blue screen appeared.
I reduced the frequency to 4.7GHz so far without any problems...
The main concern is whether these programs could still cause issues if I ran them once instead of continuously, except for the 24-hour Prime95 test...
6600k ; Z170-a ; kraken x61
There are many opinions on what to use to stress and measure OC's. I've experimented with all of them. I use Prime95, small FFTs, and run tests for an hour on the first OC and overnight for a stable result. Using small FFTs helps you check only the CPU and cache. Prime95 is designed to push the system to its limits so you can be confident it stays stable even under extreme conditions. Edit: This applies only to the CPU. I first stress my CPU, making it rock solid stable, then test my RAM separately. Independently of each other.
The reason for the BSOD at 4.8 was likely due to insufficient vcore. The main issue is the temperatures. You're running a very high OC for that chip. At stock voltage 4.7/4.8, you're right at the limit. Your temperatures...
There are many different opinions about which tools to use for stressing and measuring OC performance. I have experimented with various methods. I employ Prime95, small FFTs, and run tests for an hour initially, then overnight for what I regard as a stable result. Using small FFTs allows you to focus solely on the CPU and cache. Prime95 is designed to push the system to its limits so that stability can be confirmed even under extreme conditions without errors. Edit: This applies only to the CPU. I first stress my CPU thoroughly, ensuring it remains rock solid stable. Then I test my RAM separately. These tests are independent of each other.
The reason for the BSOD at 4.8 is likely due to insufficient vcore capacity. The main issue lies with the temperatures. You're applying a very high OC for that chip. At stock voltage, 4.7/4.8 is right at the threshold. Temperatures should stay between 30-35°C during idle, 45-65°C during gaming, and 70-80°C under stress testing using Prime95 at the specified settings. Reaching 90°C in any core is unacceptable for me. No matter the clock speed achieved during my OC runs, I cap it at 80°C. I then adjust the clock speed and voltage back to reach that target. Obviously, achieving lower than 80°C is ideal (depends on cooling) and further tuning of voltage and clockspeed is possible.
Cinebench serves its purpose as a benchmark rather than a stress-testing tool. While it's typical for basic performance checks, it shouldn't be relied upon to evaluate an OC. For determining stable voltage and maximum temperatures during CPU OC, Prime95 remains the most reliable option. All my OC attempts are based on this software.