4690k all cores running at same speed?

[monkeydelmagico]

Help guys. I've got a new 4690k and HWMonitor is showing all cores running at same speed. Asus Sabertooth Z87 mobo that I flashed to most recent bios to accept the new chip. Mild overclock by setting multi to 44 and adding 1.0 mv to boost. Speed stepping is working and all 4 cores fluctuate but they do so in lock-step.

What BIOS checkbox or Asus AI suite button do I have to push to make the cores run independently?

 

[Techhog]

You mean core parking? That's broken on i5s from what I've seen.

 

[notty22]

I believe un check MCE enhancement if you see that in bios (asus).
I only have gigabyte here, it's shown differently.

*multi-core enhancement. It's a o/c feature, that changes stock turbo behavior from normal behavior. Which is max turbo bins with only one core.

 

[VirtualLarry]

Uhm, on an Intel CPU, the frequency of all cores is in lockstep. You cannot adjust them independently. AMD is different, some of their CPUs will allow you to clock cores independently.

 

[monkeydelmagico]

Uhm, on an Intel CPU, the frequency of all cores is in lockstep. You cannot adjust them independently. AMD is different, some of their CPUs will allow you to clock cores independently.

O.K. I derped there for a minute, all better now. Thanks VL

 

[notty22]

Uhm, on an Intel CPU, the frequency of all cores is in lockstep. You cannot adjust them independently. AMD is different, some of their CPUs will allow you to clock cores independently.

You can adjust the frequency of turbo under load multiplier. K cpu's
This is intel extreme tuning utility, your bios has a similar layout under cpu features.

 

[Deders]

Have you tried Argus Monitor? The CPU Turbo tab in the performance section is usually more accurate and responsive than other frequency monitors that I've seen.

 

[VirtualLarry]

You can adjust the frequency of turbo under load multiplier. K cpu's
This is intel extreme tuning utility, your bios has a similar layout under cpu features.

And your point? Nothing I said was incorrect. The frequency for all cores in an Intel CPU runs in lockstep.

 

[notty22]

And your point? Nothing I said was incorrect. The frequency for all cores in an Intel CPU runs in lockstep.

You definition of lockstep is different than mine. one core can be running faster than the other, that's my point.

 

[VirtualLarry]

You definition of lockstep is different than mine. one core can be running faster than the other, that's my point.

No, it doesn't. Your example doesn't prove that either.

The way the turbo multipliers work, is the number of active cores. But no matter how many cores are active, all (active) cores run at the same frequency.

 

[Deders]

No, it doesn't. Your example doesn't prove that either.

The way the turbo multipliers work, is the number of active cores. But no matter how many cores are active, all (active) cores run at the same frequency.

Not necessarily, at least not on mine. Using Argus Monitor I can clearly see variations in core speeds on all 4 cores depending on the load, although if one is at it's highest then none of them will be at the lowest.

iirc AMD's chips will do as you describe.

 

[BSim500]

Not necessarily, at least not on mine. Using Argus Monitor I can clearly see variations in core speeds on all 4 cores depending on the load.

I've always wondered what the truth of this is. On the one hand I also used to believe that all Intel cores were all locked at the same freq. OTOH, in addition to mentioned Argus Monitor, differing multiplier variations can also be seen in TMonitor (edit: below is a light web browsing load. Multipliers are figures on right (16.0x is idle 1.6GHz, but see 19x & 25x on cores 1 & 3 at same time as 16x on other two cores). TMonitor rapidly polls the CPU something like 20x per second so it may "catch" things like this which other typical 1-2 polls per second CPU monitoring software don't):-

edit: Just to clarify for those unfamiliar, TMonitor measures CPU multiplier changes not CPU load.

And HWInfo:-

Are cores clocked at different rates or is it monitoring error, or something else (eg, effect of C6 states)? Anyone out there with a definitive answer?

 

[Deders]

I've always wondered what the truth of this is. On the one hand I also used to believe that all Intel cores were all locked at the same freq. OTOH, in addition to mentioned Argus Monitor, differing multiplier variations can also be seen in TMonitor (edit: below is a light web browsing load. Multipliers are figures on right (16.0x is idle 1.6GHz, but see 19x & 25x on cores 1 & 3 at same time as 16x on other two cores). TMonitor rapidly polls the CPU something like 20x per second so it may "catch" things like this which other typical 1-2 polls per second CPU monitoring software don't):-

edit: Just to clarify for those unfamiliar, TMonitor measures CPU multiplier changes not CPU load.

And HWInfo:-

Are cores clocked at different rates or is it monitoring error, or something else (eg, effect of C6 states)? Anyone out there with a definitive answer?

From what I see each core has it's own multiplier that adjusts itself depending on the load for that core. Windows will spread the load across several cores so it's rare to get just one boosting on its own.

This load is basically idle, a 1080p movie playing in the background and a couple of things moved on the desktop. Had to be quick to catch it. The orange mark shows that core 1 had just left turboboost speed a split second ago.

 

[monkeydelmagico]

I believe un check MCE enhancement if you see that in bios (asus).
I only have gigabyte here, it's shown differently.

*multi-core enhancement. It's a o/c feature, that changes stock turbo behavior from normal behavior. Which is max turbo bins with only one core.

YES! I knew I wasn't going crazy. Disabling that allows the cores to fluctuate quite a bit more depending on load instead of all 4 cores running the same frequency even if the load is only on one core.

 

[monkeydelmagico]

From what I see each core has it's own multiplier that adjusts itself depending on the load for that core. Windows will spread the load across several cores so it's rare to get just one boosting on its own.

You are the man Deders! That Argus utility is nice. HWMonitor was not as quick at updating and doesn't have a graphic display. I was having a hard time seeing what each core was doing and for how long.

With MCE enhancement disabled I'm seeing a fair amount of solitary boosting, just not for long periods of time.

With a few very brief exceptions all the cores run at the same speed though.

 

[BSim500]

From what I see each core has it's own multiplier that adjusts itself depending on the load for that core. Windows will spread the load across several cores so it's rare to get just one boosting on its own.

This load is basically idle, a 1080p movie playing in the background and a couple of things moved on the desktop. Had to be quick to catch it. The orange mark shows that core 1 had just left turboboost speed a split second ago.

That makes sense. Thanks. I see the same on mine in Argus. I guess it requires a high polling monitor to catch (ie, read CPU clocks every 50ms instead of 500-1000ms).

 

[Jovec]

Intel Haswell Docs

The following are the key features of Enhanced Intel SpeedStep Technology

• Multiple frequency and voltage points for optimal performance and power
efficiency. These operating points are known as P-states.

• Frequency selection is software controlled by writing to processor MSRs. The
voltage is optimized based on the selected frequency and the number of active processor cores.
— Once the voltage is established, the PLL locks on to the target frequency.
— All active processor cores share the same frequency and voltage. In a multi-core processor, the highest frequency P-state requested among all active cores is selected.
— Software-requested transitions are accepted at any time. If a previous transition is in progress, the new transition is deferred until the previous transition is completed.

• The processor controls voltage ramp rates internally to ensure glitch-free
transitions.

• Because there is low transition latency between P-states, a significant number of transitions per-second are possible.