Xeon E5 v3/v4 idle power consumption

[SnJ9MX]

the enclosure (read closet) tends to have a very poor airflow and I can't change this parameters. So I'd rather have a low temperature to avoid other devices to run hotter due to an elevation of the "local" temp of the closet.

The amount of heat put out by the CPU does not change based on fan speed. In fact, there will ultimately be more heat put out if you run your fans faster to keep the CPU cooler due to more electricity being used by the fans and being converted to heat due to friction (assuming the CPU is doing an equal amount of work).

Sounds like what you really need is a small vent fan for the closet itself.

 

[sergi0]

I was probably too fast to type .
At the lowest amount of noise possible, I'd rather have a lower temp in order to not increase the ambient temp and thus, the temp of other devices in the closet. That is why the X10SDV, even if it is a very nice mobo, does not fit my need as much as the Z10PA from Asus. In the later I am able to have lower temp on the same jobs running on both box.

I was wondering if the idle temp was roughly the same for all the xeon v3/v4 when idle. I got enough feedback to understand that it will be the case, as long as processors core count are in the same range. I also understand that v4 are better than v3 in power consumption when idle.

I'll probably buy a couple for xeon v4 in order to make some benchmark. The system is almost running so I will be able to compare precisely.

 

[heromode]

Here are my temps of 2x E5-2680v4 on the Z10PA, at low server load of around 5%. I use Bequiet Dark Rock slim tower coolers in a BeQuiet pure base 500 case. 2x 140mm intake fans, 1x 140mm exhaust fan, at about 1200 RPM iirc. the tower cooler fans are basically at idle, around 400 rpm.

Setting all cores to 100% with s-tui, will raise the core temps to between 60-64 degrees, and ramp up the tower cooler fans to around 600 RPM.

(2x KDE desktops are taskset pinned to package id 0, Which the GPU's running in passthrough are wired to, hence the higher temps. I can't remember which CPU is behind the other, which also raises temps abit as it gets warm air from the first tower cooler)

Without those desktops running i think all temps would be between 35 and 40 max

Package id 0:  +48.0°C  (high = +90.0°C, crit = +100.0°C)
Core 0:        +43.0°C  (high = +90.0°C, crit = +100.0°C)
Core 1:        +43.0°C  (high = +90.0°C, crit = +100.0°C)
Core 2:        +42.0°C  (high = +90.0°C, crit = +100.0°C)
Core 3:        +42.0°C  (high = +90.0°C, crit = +100.0°C)
Core 4:        +43.0°C  (high = +90.0°C, crit = +100.0°C)
Core 5:        +43.0°C  (high = +90.0°C, crit = +100.0°C)
Core 6:        +42.0°C  (high = +90.0°C, crit = +100.0°C)
Core 8:        +41.0°C  (high = +90.0°C, crit = +100.0°C)
Core 9:        +41.0°C  (high = +90.0°C, crit = +100.0°C)
Core 10:       +40.0°C  (high = +90.0°C, crit = +100.0°C)
Core 11:       +41.0°C  (high = +90.0°C, crit = +100.0°C)
Core 12:       +40.0°C  (high = +90.0°C, crit = +100.0°C)
Core 13:       +40.0°C  (high = +90.0°C, crit = +100.0°C)
Core 14:       +41.0°C  (high = +90.0°C, crit = +100.0°C)

Package id 1:  +43.0°C  (high = +90.0°C, crit = +100.0°C)
Core 0:        +35.0°C  (high = +90.0°C, crit = +100.0°C)
Core 1:        +35.0°C  (high = +90.0°C, crit = +100.0°C)
Core 2:        +35.0°C  (high = +90.0°C, crit = +100.0°C)
Core 3:        +35.0°C  (high = +90.0°C, crit = +100.0°C)
Core 4:        +35.0°C  (high = +90.0°C, crit = +100.0°C)
Core 5:        +37.0°C  (high = +90.0°C, crit = +100.0°C)
Core 6:        +35.0°C  (high = +90.0°C, crit = +100.0°C)
Core 8:        +35.0°C  (high = +90.0°C, crit = +100.0°C)
Core 9:        +35.0°C  (high = +90.0°C, crit = +100.0°C)
Core 10:       +35.0°C  (high = +90.0°C, crit = +100.0°C)
Core 11:       +34.0°C  (high = +90.0°C, crit = +100.0°C)
Core 12:       +36.0°C  (high = +90.0°C, crit = +100.0°C)
Core 13:       +36.0°C  (high = +90.0°C, crit = +100.0°C)
Core 14:       +36.0°C  (high = +90.0°C, crit = +100.0°C)

edit: adding pstate governor status:

# cpupower frequency-info
analyzing CPU 0:
  driver: intel_pstate
  CPUs which run at the same hardware frequency: 0
  CPUs which need to have their frequency coordinated by software: 0
  maximum transition latency:  Cannot determine or is not supported.
  hardware limits: 1.20 GHz - 3.30 GHz
  available cpufreq governors: performance powersave
  current policy: frequency should be within 1.20 GHz and 3.30 GHz.
                  The governor "performance" may decide which speed to use
                  within this range.
  current CPU frequency: Unable to call hardware
  current CPU frequency: 2.89 GHz (asserted by call to kernel)
  boost state support:
    Supported: yes
    Active: yes

 

[nexox]

Temperature is almost entirely irrelevant here, your computer is a space heater and the power consumption is the measure of how much heat it dissipates into the environment, any given component may be warmer or cooler based on how effectively it dissipates its heat to the air. A 50W system with a CPU temperature of 40C will warm up the closet twice as rapidly as a 25W system with a CPU temperature of 55C.

 

[bvd]

The 2650L-v4 idled at about ~6w lower than the v3 variant, at least when I'd tested it a couple years ago.

Pretty sure it was something like ~34-39w on an x10srl - I do very much miss that board, but it's living out the rest of its life at a buddy's house 1200 miles away acting as his first "real" home server (as well as an off site backup for me lol). Makes a great all rounder, 10 SATA ports off the PCH, and 7 pcie slots of various lengths.

Worth noting, I think I'd disabled some parts of the board I'd not intended to use (maybe a pcie slot or two, something like that), but can't remember whether I'd done so for the idle power test or not at this point :-/

 

[Microcosm]

When i built my Z10PA-D8 system with E5-2680v4, i did note down idle power consumption with 1CPU + 3x BeQuiet fans (probably 2x 140mm purewings2 casefans, and 1x silentwings3 towercooler fan). Nothing else, no IPMI, and it was 40W. I don't remember if i had done anything to BIOS

Just because the BMC's (IPMI) ethernet is unplugged, doesn't mean it's not running.