Extreme idle power discrepancy 3950X-5950X

[Wasmachineman_NL]

Look at this shit, this is my 3950X with PBO enabled on a ASUS Strix B550-A Gaming I use in my wipe/bench rig:

And this is the 5950X in my main, also PBO enabled:

Why is there such a ridiculous idle power discrepancy between these two computers?

Both systems are using 3600 MT/s RAM, my main rig using a built in profile for 2x16 B-Die in the C8DH BIOS and the bench rig using Teamgroup 3600 2x8 SR B-Die.

 

[Bjorn Smith]

Just a guess - it seems like from the skin of your windows that you are running windows 7 - I might be wrong - but perhaps the drivers are optimized for a more modern version of windows?

Have you tried the same in linux? To see if you can get the same difference?

 

[RolloZ170]

remember a power mismatch reporting of some motherboards, to get more power than allowed.
and compare/check with other Tools like HWinfo or AMDs own stuff.

 

[i386]

If it's really windows 7 make sure that nothing was running in the background (like that experience index thingy)

Edit: I would also use a kill-a-watt to measure power instead of relying on software

 

[mach3.2]

What's your Infinity Fabric frequency at for both Zen 2 and Zen 3 machines?

 

[Wasmachineman_NL]

What's your Infinity Fabric frequency at for both Zen 2 and Zen 3 machines?

Both at 1800. WinSAT not running on both systems, most unnecessary services also killed off/disabled in services.msc

The Eaton UPS on my main says it idles at around 105W with a single HP Z24i turned on.

 

[mach3.2]

Both at 1800.

Just an observation on my part, but it seems like the I/O die on Zen 3 does not go into a low power state at anything above 1600MHz, so the I/O die alone consumes close to 20W even on idle.

I can't tell you why Zen 2 consumes less power on idle even at 1800MHz Infinity Fabric clock though, sorry.