Here is a comparison of energy+costs between various x86 and powerpc systems. I wonder where the new efika stands. I guess i beats them all.

Easy to calculate. Efika with R9200SE, 2.5" HDD and effective power supply consumes no more than 10 W. It is:

Weekly 0.01 x 24 x 7 = 1.68 kWh (1.68 x 0.15 = $0.252)
Monthly 0.01 x 24 x 30 = 7.20 kWh (7.20 x 0.15 = $1.08)
Yearly 0.01 x 24 x 356 = 85.44 kWh (85.44 x 0.15 = $12.816)

Of course display device is not counted.

Taking Grzegorz' figures into account the Efika 5121E will use far less power than the Efika 5200B+Radeon 9250 combination - the chip, RAM and even a disk on top should clock in at somewhat less than 10W combined.

I will try and get some empirical data for you guys but since the chip is still going through revision, as is the board design based on it, it's hard to give firm facts. I'd estimate 5-7W depending on the state of the hard disk.

As for the MPC8610 Peter did some tests on his development board and got a figure of around 37W with a SATA disk, including RAM and so on. This was during a kernel compile or some other heavy activity (several tests were performed).

I am fairly sure the board has some "extraneous" development-board devices such as an FPGA and the ULi southbridge may not be as efficient as other choices we could make for a production, consumer board. Performance estimate is that at 1.3GHz it is easily 2x-3x faster than the 1GHz Pegasos G4 and 60-70% the power consumption of the ODW as reported by Arno Voelker.

This is comparable to the latest Intel mobile chips if you include their support chipsets (Northbridge, Southbridge, etc.) - remember the MPC8610 is a complete system solution lacking only the "desktop" peripherals like USB and SATA, so will also save board space and some cost.

Problem here is the PSU. My morex 60W PSU is said to have a quite high energy efficiency. But still the efika w/ Radeon 7000 and 2,5" Samsung hdd uses 15W brutto (i.e. measured at the power outlet). The PSU itself takes 6W while ATX powerdown.

See my figures here:
http://via.i-networx.de/bench_en.html

I achieved ~10W with a PicoPSU, where I could not measure anything after powerdown :)

That is the price you pay for using a transformer brick rather than a switched power supply.

Just turn the switch off at the wall and it will take 0W on power down :)

I'm a little confused about the Pegasos I. Is that really ATX shutdown or.. standby as in, the board draws power and is refreshing RAM for instant turn-on? How did you achieve that in MorphOS? :D

If it's just turning the system off, there is some discrepency here. The Pegasos I should not be drawing 6W when turned off.. it should be more on the order of less than a watt, to cover the 5V trickle for wake-on-lan style power management (not that it is implemented) and so on..

Could you resolve what "ATX shutdown" means with regards to, for instance, an ACPI power state?

It should be a switched power supply: http://www.morex.com.tw/products/produc ... p?fd_id=62
I bought that one because some ppl reported it would yield quite well, but as it seems my old laptops powerbrick/atx logic performs better. I should mearure the power uptake of the Morex brick itself though (i.e. not connected to the ATX converter logic).

I directly switch off most of my devices at the outlet (a kwh is charged 0,2 € here!).



It's G2 (S5) - that state the computer goes into when the ATX-power button (main connector pin 1&14) is pressed for 5s. I guess the wattage while in G2 state is mostly consumed by the ATX PSU itself (IIRC some no name 300W, always wantedt to replace that, but well... laziness). The Pegasos itself (i.e. the mainboard) consumes very low power in that mode of course.

Well there is a little difference between how they work, remember the PicoPSU "efficiency" ratings are for the PSU module itself and not the wall wart/brick that hangs outside the case. When you get a rating for an ATX PSU it's for the entire unit.

Just like you can get decent ~85% efficiency ATX power supplies (if you want to spend 50 EUR extra on them) you can get more efficient bricks too (if you are willing to pay 50 EUR for them rather than 5 EUR :)

But, like I said..





6W just seems rather high for a power supply only being drawn at maximum a few hundred mW on a 5V rail. Obviously with a switched-mode PSU, there are only certain levels of power it can deliver (it goes up in discrete steps) so this may be the lowest it can draw plus the inefficiency of such a load characteristic (i.e. converting 240V 50Hz mains to barely anything) showing itself. But still, 6W seems rather high..

I am surprised the brick draws that much too for the Efika, although the Efika does have a management chip to handle auto-power-on at time, it doesn't draw power except from the coin battery for a simple CMOS clock and event alarm.. so 6W "leaking" from the brick seems very high.

When my Efika is off the brick is stone cold. My VAIO laptop's brick in comparison (because it is trickle charging the battery more often than not) burns my hand if I'm not paying attention.

Marten or Neko:

Please fix that terrible URL.

Fixed and slightly edited too...:-)

Thanks!

Now I can read the thread.. :P

I measured the things on Saturday again:
The worst thing is the 12V brick, I unplugged it completely from the ATX logic and my measurement device still showed 6W. As said, I am surprised, because I ordered that particular model because it was said to be quite efficient...
My Efika (incl. the 10" LCD and the loss caused by the power brick) is typically between 28W-31W.

I just measured the power consumption of my iMX515 based EFIKA system - display device is included. It also has 4x more RAM and 2x the MHz of the original EFIKA, but no HDD, just a larger SD card. Power consumption is 9W with the bundled LCD turned on. A 2.5" USB HDD add one more Watt and an USB Ethernet adapter another one. When the LCD turns off, it spares 2-3W.

Some more results for the Efika MX here:

http://projects.powerdeveloper.org/project/imx515/732 (latest blog entry)


Johan.

Nice!