Simon Says wrote on Jul 9, 2019, 14:19:

RedEye9 wrote on Jul 9, 2019, 13:35:
To test the chipset you measure consumption at the wall.

Isn't the x570 chipset rated at 15 watts, hence the need for a fan

Der8auer measured it under different loads... https://youtu.be/qk3PD-4zPN0?t=841

He even removed the fan + crappy heatsink and went full passive by installing a very basic heatsink. He doesn't understand why they went with active cooling.

Great video, thanks. The mystery deepens.

RedEye9 wrote on Jul 10, 2019, 10:38:

Creston wrote on Jul 10, 2019, 09:56:

RedEye9 wrote on Jul 9, 2019, 13:35:
To test the chipset you measure consumption at the wall.

I will admit I am in no way an electrician, but how do you measure chipset power only if you measure at the wall?

All things being equal (cpu, gpu, ram, ssd, voltages and frequencies) with the only difference being the chipset. A kill-a-watt will do the job nicely.
You won't get the exact measure of what a component is doing but if one setup is pulling 200w and the other is at 235, it pretty much points at the different component.
Most reviewers use something similar for gpu/psu reviews, unless they've invested $100-$1000's in equipment (which some have).

Oh okay, indirectly. Makes sense. I thought you were referring somehow to measuring just that one usage at the wall.

Creston wrote on Jul 10, 2019, 09:56:

RedEye9 wrote on Jul 9, 2019, 13:35:
To test the chipset you measure consumption at the wall.

I will admit I am in no way an electrician, but how do you measure chipset power only if you measure at the wall?

All things being equal (cpu, gpu, ram, ssd, voltages and frequencies) with the only difference being the chipset. A kill-a-watt will do the job nicely.
You won't get the exact measure of what a component is doing but if one setup is pulling 200w and the other is at 235, it pretty much points at the different component.
Most reviewers use something similar for gpu/psu reviews, unless they've invested $100-$1000's in equipment (which some have).

This comment was edited on Jul 10, 2019, 10:48.

RedEye9 wrote on Jul 9, 2019, 13:35:
To test the chipset you measure consumption at the wall.

I will admit I am in no way an electrician, but how do you measure chipset power only if you measure at the wall?

RedEye9 wrote on Jul 9, 2019, 13:35:
To test the chipset you measure consumption at the wall.

Isn't the x570 chipset rated at 15 watts, hence the need for a fan

Der8auer measured it under different loads... https://youtu.be/qk3PD-4zPN0?t=841

He even removed the fan + crappy heatsink and went full passive by installing a very basic heatsink. He doesn't understand why they went with active cooling.

To test the chipset you measure consumption at the wall.

Isn't the x570 chipset rated at 15 watts, hence the need for a fan.

Nimh wrote on Jul 9, 2019, 11:43:
Interesting piece on the X570 boards. Techpowerup posted some benchmarks saying there was no appreciable difference between X470 and X570 running a Ryzen 3000 CPU. Which makes you wonder where all that power is going? (PCIe 4?)

Techpowerup probably uses a different measurement method not susceptible to this problem. The x570 chipset is fed by the 24 pins connector while the CPU is fed by the 8 or 4 pins connector(s) on top of the mobo.

Some reviewers measure the whole power usage of the PC, others use more advanced ( and precise ) methods like putting the power reader on the 8/4 pins connector, thus giving the power usage of only the CPU ( sometimes along with the RAM if the RAM is fed through the CPU connectors instead of the 24 pins ).

Some use even more advanced methods like measuring at the CPU VRM output only, those readings are unaffected by RAM, chipset and anything else on the mobo drawing power.

Interesting piece on the X570 boards. Techpowerup posted some benchmarks saying there was no appreciable difference between X470 and X570 running a Ryzen 3000 CPU. Which makes you wonder where all that power is going? (PCIe 4?)