Have I got my room temp right and does my coolers work well for me?
Have I got my room temp right and does my coolers work well for me?
I just tried messing around with some bios settings on my first overclock. My system is an i5-2500k running an Arctic Freezer 7 Pro rev2 cooler and an ASRock Z69P Pro3 motherboard with 8gb of RAM and a stock power supply. I watched a YouTube video about making 4.4ghz, but instead of the recommended values like 150 for long power consumption, I set lower ones like 125. Will my cooler handle this overclock? When I ran the Intel Burn Test in Core Temp app, the core temps were going up to 87 degrees so I stopped it. In the UserBenchmark test, my highest temp was 67 degrees using Core Temp app. Are those Intel Burn Test numbers more accurate than real-world temps? Is that level of stress better for everyday use or should I just go back down to 4.0ghz?
Use Prime95 small FFT for temperature checks. It runs at 100% continuous load using the exact same instructions that games use. Intel Benchmark Tool (IBT) uses different instructions meant for professional work, which can make things look wrong and cause a fake @ 130% CPU load. This isn't something you'd ever see in normal use or need to worry about as a baseline temp. UserBenchmark does the opposite: it adds memory usage and other stuff to the CPU load, so you have no idea what your real CPU can do, just what to expect when gaming maximums turn into a nightmare of high temps. But that's only talking about temperatures alone. You also need to run stability tests like Cinebench R20 or Asus RealBench. With those tests, you are looking for programming errors, voltage differences, and stability issues. So unless the temps get way higher than the P95 temp shown, you can ignore them because these tests are made to break your whole PC down. CPU, GPU, RAM, storage, etc., are all tested in as many different combinations it can handle. Think of it like a running test where everything gets hammered. The P95 number is flat even on the ground just to show exactly how fast you can go at full speed. The stability tests cover varied terrain with logs and trees and hills just to see when or if you trip while trying to run fast. IBT would be like running downhill, UserBenchmark would be like running in sand. Neither will give you useful results.
I tried running a prime95 test while it was already set to on. After 5 minutes, my temps got too high, like over 90C (about 93C), so I stopped the test. I've got my multiplier at 42x. Surely I shouldn't be getting these temps on an i5-2500k with a low overclock? My Vcore voltage is set to offset +0.005v and it goes up to about 1.350v during benchmarks. The VID volt reading on the coretemp says 1.4++v, but I don't know what that means. Should I change my Vcore to a fixed 1.250v in BIOS? Will that help? Is my cooler not good enough? Or is the pre-applied MX-2 paste bad? Also, can I make the fan spin faster when the CPU heats up under load? It doesn't seem to speed up even though I hear the fans running.