[SOLVED] Why is my Ryzen 5 2600X pulling 120 Watts?

[JoshC001]

I have PBO enabled on my mobo and in Ryzen Master, and I have a fairly beefy 6 heat pipe 2 fan cooler on my CPU. I use CPUID HWMonitor to pull all of my stats on my second monitor while running Cinebench on the primary display. When I run Cinebench R20, the CPU package pulls 120 Watts at around 1.35 V, clocks just over 3.9 GHz, and sits just under 70 C. I know this is at 100% full load, but from what I read the 2600X isn't supposed to pull that much power. If anyone could explain that would be great. The reason I'm going in depth with all this is for when I upgrade the CPU in the board, I know how much power the board will be able to handle in the future. Thank you!

I have linked a screenshot on my Google Drive if I sound crazy

https://drive.google.com/file/d/1OTKgmVgahEwjrO8IgrIN0APg7Sh73RxF/view?usp=sharing

 

[cdrkf]

I have PBO enabled on my mobo and in Ryzen Master, and I have a fairly beefy 6 heat pipe 2 fan cooler on my CPU. I use CPUID HWMonitor to pull all of my stats on my second monitor while running Cinebench on the primary display. When I run Cinebench R20, the CPU package pulls 120 Watts at around 1.35 V, clocks just over 3.9 GHz, and sits just under 70 C. I know this is at 100% full load, but from what I read the 2600X isn't supposed to pull that much power. If anyone could explain that would be great. The reason I'm going in depth with all this is for when I upgrade the CPU in the board, I know how much power the board will be able to handle in the future. Thank you!

I have linked a screenshot on my Google Drive if I sound crazy

https://drive.google.com/file/d/1OTKgmVgahEwjrO8IgrIN0APg7Sh73RxF/view?usp=sharing

That sounds about right to me - the 2600X is a 95W TDP (that is thermal design power)... the TDP isn't the electrical specification.

Also if you have turned PBO on, that allows the cpu to run above it's TDP limit if thermals are under control (which they will be with that cooler) - so it's going to draw more power than in stock conditions with PBO off. If you'd said 200W I would have questioned it but 120W is pretty close to standard power draw, also you have to keep in mind power readings in HWMonitor aren't 100% accurate, it could be +/- 10% of that figure easily. The only way to accurately monitor power use is to use a power meter (either at the wall socket for the whole system or you can splice into the +12V cpu power connector and ATX power cables to get a read on what the cpu is using specifically).

Edit: Link to Gamers Nexus review of 2600 / 2600X:
https://www.gamersnexus.net/hwrevie...eview-stream-benchmarks-gaming-blender/page-3
The 2600X uses 113W of power (on the 12V cpu cable) under load at stock and 130W when overclocked. PBO is a mild overclock so 120W sounds about correct imo.

 

[drea.drechsler]

.... The only way to accurately monitor power use is to use a power meter (either at the wall socket for the whole system or you can splice into the +12V cpu power connector and ATX power cables to get a read on what the cpu is using specifically).
....

You don't have to splice into the cable. Just identify all the +12V wires in the cable to the ATX-V, CPU power, connector and clip an inductive loop meter over those wires and that will measure the current through the wires. Then read the voltage on the wires with the voltmeter. The voltage multiplied by the current (many meters do that for you) will tell you power (watts) being consumed by the VRM so it includes not only the CPU's power draw but the inefficiencies of the VRM.

Reading it at-the-wall, as with a KillAWatt meter, tells you whole-system consumption. So the way to figure out how much of it belongs to a CPU during heavy loads is to do a differential reading. Watts at heavy load minus watts at low load give you a good idea of what the CPU needs for heavy processing. But this is just an approximation since the rest of the system also get more useage. The PSU, for instance, doubtless has different efficiency at low as at high load so that's wrapped up in the number too.

 

[cdrkf]

You don't have to splice into the cable. Just identify all the +12V wires in the cable to the ATX-V, CPU power, connector and clip an inductive loop meter over those wires and that will measure the current through the wires. Then read the voltage on the wires with the voltmeter. The voltage multiplied by the current (many meters do that for you) will tell you power (watts) being consumed by the VRM so it includes not only the CPU's power draw but the inefficiencies of the VRM.

Reading it at-the-wall, as with a KillAWatt meter, tells you whole-system consumption. So the way to figure out how much of it belongs to a CPU during heavy loads is to do a differential reading. Watts at heavy load minus watts at low load give you a good idea of what the CPU needs for heavy processing. But this is just an approximation since the rest of the system also get more useage. The PSU, for instance, doubtless has different efficiency at low as at high load so that's wrapped up in the number too.

I was under the impression that inductive current measurement was pretty inaccurate?

 

[drea.drechsler]

I was under the impression that inductive current measurement was pretty inaccurate?

Not as accurate is measuring voltage drop across a precision shunt resistance and then compensating for temperature rise but accurate enough for what you'll be doing. And also much safer than trying to put a meter in series with what is potentially upwards of a 40A load (many PSU's can deliver much more than that on +12V rails, actually). You'll also need a pretty beefy meter to do a series current measurement as the precision shunt resistor is inside such a meter.

Kill-A-Watt's use an inductive loop to measure current inside those tiny little cases.