Question are my core temps safe and is my cpu cooler good enough ?

[malibu200786]

ive just messed around with bios settings for my first overclock .
my system is . i5 2500k . arctic freezer 7 pro rev2 cooler
asrock z68 pro3 motherboard
8gb ram
stock psu

i watched a video on doing 4.4ghz on youtube , i followed what he did but put in slightly lower values like
125 for long power consumption instead of 150 etc .
will my cooler be good enough for this overclock ?

when i did the intelburn test the core temps where going up to 87 degrees in coretemp app , so i stopped the test .
when i did the userbenchmark test my highest temp was 67 degrees in coretemp app
are the intelburn test results not the real world temps ? is that more stress than the cpu will get in everyday use ?

should i revert back to 4.0ghz ?

 

[Karadjgne]

Use Prime95 small fft for temp testing. It's a continuous 100% load using the same instruction sets that games use.

IBT uses some different instructions, ones geared towards professional content creation etc so can result in squewed results that can hit an artificial @ 130% cpu load. Not something you'd ever run into or use, so not even a good baseline temp to get your usage from.

Userbenchmark does the opposite, it adds in memory usage and others to cpu load, so you end up with no idea what your cpu can really do and just what to expect as a gaming maximum nightmare temp.

But that's just temps alone. You'll also need to run a stability test, like cinebench r.20 or Asus RealBench. With those tests you are after programming errors, voltage discrepancies, stability etc. So unless temps far exceed the p95 temp, you can ignore them, as the tests are designed to brutalize your entire pc. Cpu, gpu, ram, storage, etc, in as many combinations as it can.

Think of it as a running test. P95 is flat, even ground, just to see exactly how fast you can run flat out. The stability tests are over varied territory, logs, trees, bushes, hills just to see when or if you trip when trying to run fast. IBT would be like running downhill, userbenchmark running in sand. Neither will get you useful results.

 

[malibu200786]

Use Prime95 small fft for temp testing. It's a continuous 100% load using the same instruction sets that games use.

IBT uses some different instructions, ones geared towards professional content creation etc so can result in squewed results that can hit an artificial @ 130% cpu load. Not something you'd ever run into or use, so not even a good baseline temp to get your usage from.

Userbenchmark does the opposite, it adds in memory usage and others to cpu load, so you end up with no idea what your cpu can really do and just what to expect as a gaming maximum nightmare temp.

But that's just temps alone. You'll also need to run a stability test, like cinebench r.20 or Asus RealBench. With those tests you are after programming errors, voltage discrepancies, stability etc. So unless temps far exceed the p95 temp, you can ignore them, as the tests are designed to brutalize your entire pc. Cpu, gpu, ram, storage, etc, in as many combinations as it can.

Think of it as a running test. P95 is flat, even ground, just to see exactly how fast you can run flat out. The stability tests are over varied territory, logs, trees, bushes, hills just to see when or if you trip when trying to run fast. IBT would be like running downhill, userbenchmark running in sand. Neither will get you useful results.

I tried a prime95 test on the setting it's already on when starting a test .
After 5 minutes the temps where above 90oC like 93oC , so I stopped the test .

I've got the multiplier on 42x
Surely I shouldn't be getting them temps on i5 2500k at a low overclock ?
The Vcore voltage I put to offset +0.005v
And is going upto 1.350v ish when doing benchmarks .
Also the VID volt reading on coretemp is saying 1.4++v I don't know what this reading is ?

Should I set my Vcore to 1.250v fixed on bios ? Will this help ?

Is my cooler not good enough ? Or is it the pre applied MX-2 paste which is rubbish?

Also can I make the fan spin faster , I don't hear it speed up when CPU under load