i7 8700k delid wattage

[AkSo2504]

Hello,

i delided my i7 8700k with great suces in temperatures, but msi afterburner now all the time shows that the cou is using only 3-6 watts.

Deos anyone know why?

[KarathKasun]

Probably because it calculates wattage off of temps and input voltage.  You drastically changed the CPUs thermal resistance so this value is now incorrect.

[Ebony Falcon]

46 minutes ago, AkSo2504 said:

Hello,

i delided my i7 8700k with great suces in temperatures, but msi afterburner now all the time shows that the cou is using only 3-6 watts.

Deos anyone know why?

Yeah mine is same and I don't no why 

[AkSo2504]

22 minutes ago, Ebony Falcon said:

Yeah mine is same and I don't no why 

ok good to know, its probably like karath already said.

[Falkentyne]

4 hours ago, AkSo2504 said:

Hello,

i delided my i7 8700k with great suces in temperatures, but msi afterburner now all the time shows that the cou is using only 3-6 watts.

Deos anyone know why?

All the answers in this thread are incorrect.

There is no such thing as "thermal resistance" in relation to the CPU in this context.  Resistance is measured in mOhms, and is a function of loadline.

If you're referring to thermal paste thermal resistance, that has ZERO influence on CPU VID, which is a fixed preset based on CPU Core+Cache frequency multipliers, biased by AC Loadline (Sent to the VRM on auto voltages), then droop biased independently by DC Loadline.

 

What are you trying to read? CPU Package Power?

Or are you trying to read something else from the VRM, like Power POUT (Some Asrock, MSI and Gigabyte boards have access to this)?

 

CPU package power is exactly VID * Amps.

Power POUT is exactly VR VOUT * Amps.

 

Post your CPU Core VID (for any active core) at idle and full load.

Are you using C-states or downclocking/power saving?

[AkSo2504]

17 hours ago, Falkentyne said:

All the answers in this thread are incorrect.

There is no such thing as "thermal resistance" in relation to the CPU in this context.  Resistance is measured in mOhms, and is a function of loadline.

If you're referring to thermal paste thermal resistance, that has ZERO influence on CPU VID, which is a fixed preset based on CPU Core+Cache frequency multipliers, biased by AC Loadline (Sent to the VRM on auto voltages), then droop biased independently by DC Loadline.

 

What are you trying to read? CPU Package Power?

Or are you trying to read something else from the VRM, like Power POUT (Some Asrock, MSI and Gigabyte boards have access to this)?

 

CPU package power is exactly VID * Amps.

Power POUT is exactly VR VOUT * Amps.

 

Post your CPU Core VID (for any active core) at idle and full load.

Are you using C-states or downclocking/power saving?

its only when im reading cpu package power, cpu power itself looks normal.

My voltage is 1.232v at idle and 1.216v at load.

Im not using c-states, the only thing i changed is that i delided the cpu, but ebony falcon says the same, so im not alone.

[Falkentyne]

Could be a bug.

CPU Package Power can be influenced by IMON slope/offset too.

[KarathKasun]

On 5/31/2019 at 5:19 PM, Falkentyne said:

All the answers in this thread are incorrect.

There is no such thing as "thermal resistance" in relation to the CPU in this context.  Resistance is measured in mOhms, and is a function of loadline.

If you're referring to thermal paste thermal resistance, that has ZERO influence on CPU VID, which is a fixed preset based on CPU Core+Cache frequency multipliers, biased by AC Loadline (Sent to the VRM on auto voltages), then droop biased independently by DC Loadline.

 

What are you trying to read? CPU Package Power?

Or are you trying to read something else from the VRM, like Power POUT (Some Asrock, MSI and Gigabyte boards have access to this)?

 

CPU package power is exactly VID * Amps.

Power POUT is exactly VR VOUT * Amps.

 

Post your CPU Core VID (for any active core) at idle and full load.

Are you using C-states or downclocking/power saving?

Actually, there is.  De-lid reduces the thermal resistance between the die and cooling solution, thermal resistance is expressed in W/mK.  AFAIK, the on die "power" reading is calibrated to the package (because there is no sensor for power consumption on the CPU) and modifying the package breaks this calibration causing garbage output.

[Falkentyne]

3 hours ago, KarathKasun said:

Actually, there is.  De-lid reduces the thermal resistance between the die and cooling solution, thermal resistance is expressed in W/mK.  AFAIK, the on die "power" reading is calibrated to the package (because there is no sensor for power consumption on the CPU) and modifying the package breaks this calibration causing garbage output.

I'm sorry but this makes absolutely zero sense, and I'm calling bollocks on your post.

 

Package Power is VID * Amps.

 

Amps has nothing to do with thermal resistance, period.

 

VID is hardwired into the CPU based on core and cache ratio, and then boosted by AC Loadline (up to max VID, 1.52v).

This value is sent to the VRM as target voltage after AC Loadline bias on auto/adaptive voltages only (more biasing can occur based on adaptive).  This value is then drooped by VRM Loadline (aka CPU Loadline Calibration value) down to a safe value, as long as Loadline calibration is left at default (1.6 mOhms for 8 core processors).  

 

DC Loadline drops this VID (VRM ignores this completely) based on DC Loadline mOhms, as DC loadline is used for power reporting only.

CPU Package Power is based on this final result (VID * Amps) and biased by IMON SLOPE and IMON OFFSET.

Plenty of people go direct die cooling and still have proper results also, without an IHS.

[KarathKasun]

6 hours ago, Falkentyne said:

I'm sorry but this makes absolutely zero sense, and I'm calling bollocks on your post.

 

Package Power is VID * Amps.

 

Amps has nothing to do with thermal resistance, period.

 

VID is hardwired into the CPU based on core and cache ratio, and then boosted by AC Loadline (up to max VID, 1.52v).

This value is sent to the VRM as target voltage after AC Loadline bias on auto/adaptive voltages only (more biasing can occur based on adaptive).  This value is then drooped by VRM Loadline (aka CPU Loadline Calibration value) down to a safe value, as long as Loadline calibration is left at default (1.6 mOhms for 8 core processors).  

 

DC Loadline drops this VID (VRM ignores this completely) based on DC Loadline mOhms, as DC loadline is used for power reporting only.

CPU Package Power is based on this final result (VID * Amps) and biased by IMON SLOPE and IMON OFFSET.

Plenty of people go direct die cooling and still have proper results also, without an IHS.

Amp draw has a direct effect on thermal load, and with a known thermal resistance you can work the equation backwards and get power from volts and temperature.  The resulting power number would be "in the ballpark", which seems to be about as accurate as most motherboard algos.

 

Regardless, it seems different boards employ different strategies to calculate this value and report it.  The same chip in two different boards can have wildly different "board reported power".  Enabling any of the non-stock LLC settings could also have a massive impact on this reported value, and may in fact mis-report the value to fool the intel power management into self overclocking.

 

The final question would be, did any of the users change LLC or any "auto-oc" options after the de-lid?

[Falkentyne]

1 hour ago, KarathKasun said:

Amp draw has a direct effect on thermal load, and with a known thermal resistance you can work the equation backwards and get power from volts and temperature.  The resulting power number would be "in the ballpark", which seems to be about as accurate as most motherboard algos.

 

Regardless, it seems different boards employ different strategies to calculate this value and report it.  The same chip in two different boards can have wildly different "board reported power".  Enabling any of the non-stock LLC settings could also have a massive impact on this reported value, and may in fact mis-report the value to fool the intel power management into self overclocking.

 

The final question would be, did any of the users change LLC or any "auto-oc" options after the de-lid?

Power draw does go up with temps, yes, so you are correct there.  Amps does go up.

Although I wouldn't call that thermal resistance in the same way you were relating that to the raw CPU package power.

Remember there is also thermal velocity boost also, where starting at 100C, the VID drops by 1.5mv (0.0015v) every 1C temp drop.

(Asus lets you enable or disable this), or if it's not an option in the bios, it can be enabled via a MSR in RWEverything.

Don't confuse that with the laptop version of TVB which is ratio overclocking below 50C (good luck).

 

Now for the current emphasis on totally stock perf of the i9’s by the review sites, all the attention is on TDP but that’s just a gnat compared to the camel swallowed. NO site actually talked about and examined the latest feature of the i9, Thermal Velocity Boost TVB. By default Intel enables this but I see that only Asus boards enable this at defaults. The other boards I tested have this disabled even at defaults. 

What this does is it reduces voltage guardbands depending on core temp. Traditionally, the voltage request by the proc is always based on worst case scenario TJMAX, meaning the voltage the proc thinks it needs for the frequency when temp is 100c. It is well-known that the cooler the chip runs, the lesser the voltage needed. Therefore TVB is opportunistically reducing power and temps. The behavior is quite linear and I observed the following on several samples. 

TVB takes effect from 40~50x on 99k and 40 to 49x on 97k and 40 to 47x on 96k, simply 40x to single core boost ratio. The V/temp curve runs from 0c to 100c. For example 150mv delta between 100c and 0c for 50x, meaning every 1C drop from 100c VID requested will reduce by 1.5mv. The reduction is smaller as you go down to 49x, the smaller the ratio the smaller the reduction, and below 40x you get no reduction. This is good for most people running stock. You can try this yourself by noting the VID idle, and then unplug your water pump and let the core temp rise slowly, noting down the correlated temp/VID, and see what i'm talking about.