Check if your main temperatures are normal and if your CPU cooler performs adequately.
Check if your main temperatures are normal and if your CPU cooler performs adequately.
i just experimented with bios settings for my first overclock.
my system is an i5 2500k, arctic freezer 7 pro rev2 cooler, asrock z68 pro3 motherboard, and 8gb ram with a stock psu.
i watched a youtube video about reaching 4.4ghz, but i adjusted the values slightly lower, like 125 for long power use instead of 150.
will my cooler handle this overclock?
during intelburn test, core temps reached up to 87 degrees in coretemp app, so i stopped the test early.
the highest temp recorded during userbenchmark was 67 degrees in coretemp app.
are the intelburn results accurate or just a stress test? is it more demanding than everyday use?
should i go back to 4.0ghz?
Prime95 small fft is suitable for quick temperature checks. It applies a full load with identical instruction sets used in gaming, making it a 100% continuous test. IBT offers different instructions aimed at professional workflows, which may lead to skewed outcomes and artificial high CPU usage. These aren't typical benchmarks for real-world performance evaluation.
Userbenchmark incorporates additional factors like memory consumption and other metrics, obscuring the actual CPU capabilities and making it hard to gauge expected gaming performance.
Temperature readings alone are insufficient. It's essential to perform stability assessments such as Cinebench R20 or Asus RealBench. These tests evaluate programming issues, voltage variations, and overall system stability across diverse conditions like logs, trees, hills, etc. Unless temperatures surpass the P95 mark, they shouldn't be relied upon for meaningful insights.
CPU, GPU, RAM, storage, and other components should be tested in various configurations to simulate real usage scenarios. Consider it a comprehensive stress test—P95 provides a consistent baseline, while stability tests reveal hidden issues across different environments. IBT resembles a downhill run, whereas Userbenchmark feels like running in sand; neither delivers actionable results.
I ran a prime95 test and noticed the temperatures were already at 90°C when starting. After about 5 minutes, they reached around 93°C, so I ended the test early. The multiplier is set to 42x. It seems I shouldn't expect such high temps on an i5 2500k with low overclocking? I'm using a Vcore offset of +0.005v and planning to go up to 1.350v during benchmarks. The Vcore voltage reading in core temp is showing 1.4v plus, which is unclear. Should I lock the Vcore at 1.250v in BIOS? Would that help? Am I using a cooler that's insufficient? Or is the pre-applied MX-2 paste not good quality? Also, can I increase the fan speed, but it doesn't seem to respond when the CPU is under load?