I’ve checked power consumption of my rig today, and was a bit buffed:
106W while idle on desktop (Fedora 31/Gnome)
160W with 8 maxed out threads.
Is it normal for Ryzen 1600?
MB: Gigabyte AX370-gaming 5 (bios F42e);
PSU: Corsair AX 860;
no RGB besides MB.
GPU: RX580
2 SSD drives.
Measured from the wall socket.
100 watts at idle seems steep! How did you measure it at the wall socket?
Measured with power counting device
Not exactly this one, but with almost identical layout.
I’ve checked it already with device with known consumption, seems ok.
Yea, that should be pretty accurate. I have KillaWatt, will check it out on my Ryzen 2600 tonight, hopefully I’ll remember 
Thank you, I had my rig overclocked (was about 30W more under 66% load), the numbers I mentioned in first post are without OC with optimized settings (besides RAM: 2666->3200).
Well what you measure from the wall is not really what the cpu is using.
You are basically measuring total system power,
which means the power draw for the entire system.
This also varies with psu efficiency ratings etc. 
If you want to measure how much power your cpu only is consuming in a given situation,
then that is a totally different measure and calculation procedure.
Checked mine using OpenCorsairLink on Fedora 31. At idle it reports 108W with an R7 1700, AsRock X370 Taichi, and two RX 580s with around 208W maxing out all 16 threads while compiling the Linux Kernel.
Raw data for the results
Idle
michael@fedora ~/git/OpenCorsairLink(testing) » sudo ./OpenCorsairLink.elf --device 0
Dev=0, CorsairLink Device Found: RM1000i!
Vendor: CORSAIR
Product: RM1000i
Firmware: NA
Temperature 0: 50.16 C
Temperature 1: 44.02 C
Powered: 109452699 (1266d. 19h)
Uptime: 19899 (0d. 5h)
Supply Voltage: 230.00 V
Total Watts: 108.00 W
Output 12v:
Voltage 12.16 V
Amps 7.00 A
Watts 84.00 W
Output 5v:
Voltage 5.06 V
Amps 3.88 A
Watts 19.00 W
Output 3.3v:
Voltage 3.34 V
Amps 2.00 A
Watts 7.00 W
cpu load
Dev=0, CorsairLink Device Found: RM1000i!
Vendor: CORSAIR
Product: RM1000i
Firmware: NA
Temperature 0: 51.18 C
Temperature 1: 44.02 C
Powered: 109453422 (1266d. 19h)
Uptime: 20622 (0d. 5h)
Supply Voltage: 230.00 V
Total Watts: 208.00 W
Output 12v:
Voltage 12.12 V 2
Amps 15.00 A
Watts 180.00 W
Output 5v:
Voltage 5.03 V 6
Amps 4.75 A
Watts 24.00 W
Output 3.3v:
Voltage 3.34 .34
Amps 1.75 A
Watts 6.00 W
cpu load + furmark on one RX580 for the lulz
Dev=0, CorsairLink Device Found: RM1000i!
Vendor: CORSAIR
Product: RM1000i
Firmware: NA
Temperature 0: 52.72 C
Temperature 1: 44.02 C
Powered: 109453502 (1266d. 19h)
Uptime: 20702 (0d. 5h)
Supply Voltage: 230.00 V
Total Watts: 404.00 W
Output 12v:
Voltage 12.12 V 2
Amps 31.00 A
Watts 376.00 W
Output 5v:
Voltage 5.03 V 6
Amps 4.88 A
Watts 24.00 W
Output 3.3v:
Voltage 3.34 .34
Amps 2.25 A
Watts 7.00 W
1 Like
So you also sport 100W idle with Ryzen - it’s quite a lot.
On Guru3d.com they got 50W with “normal” setup’ (hasn’t seen any fans there, but 1 fan is about 1,5W max and they had even bigger PSU).
Thank you for measuring.
1 Like
You are technically right
But there is still quite a gap between Intel and AMD systems.
Can’t find my kill-a-watt, so I’m a bit delayed on this party.
1 Like
I think that a combination of Linux being more “aggresive” on power and the 240V gave a different number to Guru3D than your (240V allows for better PSU efficiency so at idle they’re consuming less power).
Maybe if you play with the CPU scheduler and some other tools you’ll be able to take the power consumption down.
Also I don’t know if Linux does this but Windows (or the AMD driver on Windows to be more precise), to save power, even drops the PCIe link to 1x if you’re not doing anything intensive with the GPU.
I do use 240V.
You are right, windows is a bit less power hungry (90W idle).
Not sure about my monitor’s refresh rate on Linux, but on Windows changing it from 144Hz to 60Hz yields about 20W savings.
Could be, it was not the most brilliant idea to put 5 fans in the case, but I had 2 GPUs and it was necessary.
Changing scheduler - hmm, it means compiling kernel if I got it right - a bit too much hustle for probably low yields.
Thank you all for your input!
Unless you’re running some weird distro, you can change scheduler on the fly from the command line.
Oh, didn’t know that sorry.
That’s absolutely one more reason why your system uses more power, as you saw.
If they’re not running at full blast and they’re not 120mm screaming banshees you should be using 3-4W more at idle.
You’re welcome, happy to have brought up useful ideas!
1 Like
This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.