It's the other way around, voltage dictates how much current the CPU draws. If you give it more voltage, it draws more current, thus increasing the power that the CPU uses.

computerguy08 wrote on 2020-07-03, 11:18:

It's the other way around, voltage dictates how much current the CPU draws. If you give it more voltage, it draws more current, thus increasing the power that the CPU uses.

P=U*I, so for the same power, when decreasing the voltage, the current must increase.

Oetker wrote on 2020-07-03, 11:32:

P=U*I, so for the same power, when decreasing the voltage, the current must increase.

P is not a constant, it will adapt to the level of voltage. Why do you think that you can run LEDs from a CR2032 battery just fine, but you need resistors if you plan to run them at 5 (or even 12) volts ?

On my Asus P3B-F, a 1,4 Tualatin works with an Upgradeware Slot-T (no active VRM), but the 2 MOSFETs in the SLOT1 area heat up significantly. I wouldn't expect yours to behave otherwise.

I've added two GPU memory heat sinks on these, and while benching, I had a 12 cm fan blowing on the whole area. Even at an FSB of 150, that kept temperatures on these components at an acceptable level.

flupke11 wrote on 2020-07-03, 11:38:

On my Asus P3B-F, a 1,4 Tualatin works with an Upgradeware Slot-T (no active VRM), but the 2 MOSFETs in the SLOT1 area heat up significantly. I wouldn't expect yours to behave otherwise.

I've added two GPU memory heat sinks on these, and while benching, I had a 12 cm fan blowing on the whole area. Even at an FSB of 150, that kept temperatures on these components at an acceptable level.

Interesting, I plan on running at FSB100, but I doubt that will result in a 25% power decrease.

computerguy08 wrote on 2020-07-03, 11:37:

Oetker wrote on 2020-07-03, 11:32:

P=U*I, so for the same power, when decreasing the voltage, the current must increase.

P is not a constant, it will adapt to the level of voltage. Why do you think that you can run LEDs from a CR2032 battery just fine, but you need resistors if you plan to run them at 5 (or even 12) volts ?

It is a constant if Intel says that for a certain CPU at its stock voltage, it consumes X power. Of course its power consumption will decrease if I decrease its voltage relative to that. However I'm comparing two different CPUs with the same TDP running at different voltages.

Oetker wrote on 2020-07-03, 11:45:

However I'm comparing two different CPUs with the same TDP running at different voltages.

That slipped trough my fingers. If you have two CPUs with the same TDP, but different process nodes (thus different voltages), then you are indeed correct.

It seems like sometimes I'm not paying enough attention to the details 😁

I don't know how to derive the VRM capabilities from looking at the MOSFET specifications, but you might be able to get a few insights from the following aspects:

• The most current-hungry CPU Asus lists as compatible for the P3B-F is a Celeron 1100.
• According to Intel datasheets, this CPU draws a core current (ICC_Core) of up to 22.6 A. (This should be the main power consumer of the CPU.)
• You can look up the ratings for Tualatin-based processors in other datasheets: Celeron, Pentium III, Pentium III-S

Edit: Sorry, I got the impression that you intended to run the processor on a P3B-F, but that wasn't you writing that. In any case, should you find the compatibility lists for your mainboard, you might be able to make a similar deduction.