Benchmark Programs

[Torn]

6600k Oc'd to 4.5Ghz.

Any idea why OCCT, Prime 95 and Intel Burn Test will only ramp up my cpu to 3.3GHZ when Cinebench will push it to 4.5?

[porina]

What cooling are you using? Have you monitored CPU temperatures?

[suchamoneypit]

AIDA64 works fine testing my OC.

[Torn]

With Aida64, core clocks jump around from 4.0.to 4.3 to 4.5.

Temp average for most benchmarks is under 60c. max I've seen on a single core was 65C.

[porina]

So not thermal throttling then.

 

Only other thing I can think of is current limiting, but in my only experience hitting that, it is all or nothing and I didn't see intermediate frequencies. Unless there are multiple things going on, such as variable turbo combined with the current limit?

 

I'm most familiar with Prime95, and it would be interesting to see what that does if you manually tell it to stress 1 core only, then 2, then 4. I would also suggest running hwinfo as that can show various flags for limiting which might also give a clue.

[Torn]

I'm running CPUID HWMonitor. That's where I'm seeing the core clocks jump around.

If I run a Cinebench Cpu test with Prime running, it will spool up the core clock to 4.5. Soon as Cinebench is finished rendering the clock drops back to 3.3.

 

[Torn]

And to answer your question, I'm using a Noctua NH-U14S

[porina]

HWinfo is similar to HWmonitor, but HWinfo has a bit more stuff in it. The status flags attached are what I'm curious about as they may give a clue to what's going on.

[Torn]

Ok so if I tell OCCT to test a single thread, utilization is 100 on core 0, minimal on the rest. (0 to 3%) But, all core clocks go to 4.5.

2 threads, utilization core 0 ( 30to 50%), core 1 (5 to 30%). core 2 (40 to 70%) and core 3 (70 to 96%) Clocks jumping around from 3.8 to 4.5

3 threads. higher utilization, just as random. Clocks 3.2 to 4.2

4 threads, 100% clocks from 3.3 to 4.3

[Torn]

[porina]

The upper one with a "Yes" in it is the same one I saw when the CPU was hitting current limit. I found if I relaxed the limit to 130A from default 120A it went away and didn't throttle back, but I don't know if that might vary depending on CPU/mobo. Setting will be buried somewhere in the bios, or you can set it through Intel XTU.

 

I found it interesting I only got that on one of my i7-6700k CPUs, and it didn't happen on my i5-6600k either. So this is likely due to some variation in the CPUs.

[Torn]

I've rolled back to 4.3Ghz oc, with 1.25v,  still the same issue. I'll go see what I can did up in the for current limit.

[Torn]

Same issue at 130A, how high is too high I'm wondering.

[Torn]

Seems to be an issue with MSI boards, but it's hard to find any solid information on safe current limits.

[porina]

Which mobo do you have? I saw the effect in an MSI Gaming Pro. I have two of those, and it seemed to follow the CPU, not the motherboard.

[Torn]

MSI Z170A PC MATE