mgiammarco wrote:
Zolishoru wrote:
What about an
Intel S1200KPR +
Pentium G2120 combination(yes, it does support ECC)? Idle power isn't too far from the atom, but with way more processing power, and as price should be close.
Are you sure about idle power? Where I can see some data?
I've done some measurements some time ago, and
published the results, reproduced below:
Quote:
Hardware configuration:
MB: Intel S1200KPR
CPU: Intel Pentium G2120
CPU heatsink: Scythe Mini Ninja Rev. B, no fan attached
RAM: Kingston KVR133D3E9SK2/16G, 16(2X8)GB DDR3-1333 ECC
SSD: Patriot Warp V2 32GB
KB: Logitech Illuminated, with lights turned off
Mouse: Gigabyte GM-M6880
Fans: Noctua NF-P12-1300(with ULN) and NF-P14-FLX(no adapters), both connected to the MB
DC/DC converter: WinMate DD-24AX
Power brick: Acer, 19V/130W, class IV efficiency
OS: Ubuntu 12.10 64bit
Multimeters: Tektronix DMM870(current), Fluke 75 Series II(voltage)
Results:
1.
DC/DC converter, alone, off:
19.6V*0.01A=0.2W
DC/DC converter, alone, on:
19.6V*0.2A=3.9W
2.
16GB RAM(2modules):
System on, after 10 minutes idle on the desktop:
19.5V*1.01A=19.7W
System on, after 10 minutes idle on the desktop without KB&M:
19.5V*0.97A=18.9W
System running Prime95, 2 treads, in-place large FFT:
19.3V*2.29A=44.2W
System running Prime95, 2 treads, in-place large FFT, no KB&M:
19.3V*2.26A=43.6W
System running Prime95, 2 treads, blend test:
19.3V*2.33A=45W
System running Prime95, 2 treads, blend test, no KB&M:
19.3V*2.3V=44.4W
3.
8GB RAM(1 module), KB&M attached:
System on, after 10 minutes idle on the desktop:
19.5V*0.95A=18.5W
System running Prime95, 2 treads, blend test:
19.3V*2.15A=41.5W
4.
Currently used configuration:
MB: Intel S120KPR
CPU: Intel Pentium G2120
RAM: Kingston KVR133D3E9SK2/16G, 16(2X8)GB DDR3-1333 ECC
HBA: IBM M1015, flashed in IT mode
HDD: Samsung 640GB, 7200 RPM
Boot device: SanDisk Cruzer Fit 16GB USB stick
No KB&M
Booted up & idled for 15min in FreeNAS 8.3, no power management:
19.3V*1.95A=37.6W
Notes:
1.At all tests, a network connection was active(1GbE).
2.All current measurements are done between the power brick and the DC/DC converter.
3.All voltage measurements are done at the input of the DC/DC converter, after the current multimeter.
4.Ambient temperature during the tests was 22C, the CPU idled around 23-24C, and under full load was between 45-50C
5.Multimeters are NOT calibrated, because for home use make no sense third party certifications; however, on yearly basis I compare both multimeters against a calibrated HP 34401A multimeter, to which I have access at my workplace; so far, both are within the original factory specs. Last check was in July 2012.
The results are similar to the
SPCR G2120 tests(different MB).