Power useage reports - how many watts does FAH add?

MixManSC

MixManSC

║▌║█║▌│║▌║▌█
Staff member
Joined
Aug 12, 2004
Messages
8,374
I've not seen much in the way of this so I thought I'd start some testing on power use of my machines. I'm testing with a "Kill a Watt - KAW" (about $20 on Amazon.com) which shows your power use. My method is going to be testing with the tower only connected through the KAW and I'm going to report idle power use and power use with FAH running. I'm really just trying to get a better idea of how much juice I'm really using by folding and what my boxes are using when sitting idle. I'll be updating this regularly over the next couple of weeks as I get to each machine and I'll also be posting the specs of each machine along with OS and FAH client(s) that are being run.

If any of you other folders happen to have a KAW it would be great if you added your results as well. In the end I'll compile all the results into a spreadsheet. Maybe some of you might find this useful, maybe not. Either way I think it will be an interesting experiment. The results are for the tower only, no monitor, UPS, cable modems, etc. I've found so far that I need to let FAH run for a few minutes to get fully started up to get valid numbers. First number is watts, second number is amps. Idle results are just that - sitting at the desktop with no scans, defrag, screen savers, etc. or any other activity happening. Stuff is running but not active - system idle process at 95% or higher.


First machine is my main home rig.

Socket 478 P4 3.2Ghz 512K Northwood
Abit IS7 Mobo
2GB Corsair XMS PC3200 (2x1GB)
3 300GB Hard Drives
2 Optical Drives
600W Ultra Power Supply
Nvidia 6800GT AGP Video
Windows Vista
Single FAH504 Console Client

Idle - 122 W - 1.65 A
FAH Running - 178 W - 2.35 A

Results on this machine show a roughly 55W increase in power use by running FAH.
 
You'll definately want to test different machines. The old P4 netburst chips were power-hogs and could double as a frying-pan cooking surface. Would be nice to get an array of Pentium D, A64 and C2D power usage results.
 
I don't have any results for chips at idle, but I do have results at full power. Some of the machines can be tested again, but most have been upgraded.
 
My boxen I've checked are .....

Eeyore Boxen ..........

C2D E6600 @ 2.75 Ghz.
2 Gigs memory, Raptor hard drive, 4x120mm cooling fans.
Running the SMP client.
~1400 PpD for ~150 watts used.

Kanga, Roo, Rabbit, Owl & Pooh Boxen .........

4200+ X2's with X1950xtx Vid cards.
2 Gigs memory, SATA hard drive, 3x120mm cooling fans.
Running GPU & CPU clients.
~950 PpD for ~250 watts used.

Piglet Boxen ...............
Dual 250 Opteron's
2 Gigs memory, SCSI hard drive, 4x120mm cooling fans.
Running the SMP client.
~900 PpD for ~350 watts used.

Tigger Boxen .............
Dual 285 Opterons with Dual X1950xtx vidio cards.
8 Gigs memory, 3 Raptor hard drives, 5x120 mm cooling fans.
Only running the SMP client at the moment.
~1800 PpD for ~450 watts used.

If I use the SMP client on my X2's then each drop ~200 PpD & ~100 watts power.
If I'd known how well the C2D's folded I would of held off my upgrade from MP's to X2's in favour of them.
Piglet boxen is the next one to be upgraded, probably to 5355's,as not enough PpD for the power used at the moment.

Edit, No idea how much power each uses idle as are never idle ..... :p

Luck .......... :D
 
I don't have the exact figures anymore (lost them), but as I recall:

AMD XP 3800 OC'd to 2.4ghz
2 GB memory,
Nvidia 6800 Ultra vid card
220w. Knock off about 40w for the video if you put in a cheapo PCI-card.

AMD XP 3800 OC'd to 2.4ghz + 1 ATI X1900XTX video card, running 1 CPU client and 1 GPU client.
1 GB memory
PCI video card
250w

Intel E6400 OC'd to 3.2ghz
1 GB memory
PCI video card
153w

AMD A64 3400 Socket 754
512MB memory
PCI video card
130w

AMD A64 3000 dialed down to 1.4ghz, less than 1v CPU.
Running GPU client on a X1900XTX
1 GB memory
170w

These numbers were with typical 70% efficient power supplies; I replaced them with 83-84% supplies but haven't re-measured since then.
 
wow, some of the figures i am seeing here are odd. lets try mine.

my amd dedicated boxen all have tbred's or barton's o/c'ed to 2.3ghz with the exception of two that run at 2ghz and 2.2ghz. they all have 2 sticks of pc3200 ram at 2.6v's and they all have very low-end, low power agp cards. they average at 120watts at full load.

my intel dedicated boxen have different overclocks, have 2 sticks of pc3200 ram at 2.6v's and have the same agp cards as my amd boxen and they average at 200watts at full load.

my server with an o/c'ed p4c, 2gigs of ram, 9700pro, and 16hd's tops out at 500watts at full load :D

my main gaming rig with an o/c'ed FX60 @ 3ghz, 2gigs of ram, X1950pro, 3 hd's tops out at 300watts at full load (both cores and gpu)
 
Bill Clo said:
I don't have the exact figures anymore (lost them), but as I recall:

AMD XP 3800 OC'd to 2.4ghz
2 GB memory,
Nvidia 6800 Ultra vid card
220w. Knock off about 40w for the video if you put in a cheapo PCI-card.

AMD XP 3800 OC'd to 2.4ghz + 1 ATI X1900XTX video card, running 1 CPU client and 1 GPU client.
1 GB memory
PCI video card
250w

Intel E6400 OC'd to 3.2ghz
1 GB memory
PCI video card
153w

AMD A64 3400 Socket 754
512MB memory
PCI video card
130w

AMD A64 3000 dialed down to 1.4ghz, less than 1v CPU.
Running GPU client on a X1900XTX
1 GB memory
170w

These numbers were with typical 70% efficient power supplies; I replaced them with 83-84% supplies but haven't re-measured since then.
Click to expand...

Yup, if you recall and I know you do because you remembered your numbers, we did this very informally twice before.

A spread sheet is a probably a good idea this time for future comparisons.

The important info is KWH full load and no load basic system specs and these days, exactly which PSU you are using.

Although every KWH costs money my findings are folding adds very little to the overall cost of leaving the computer running 24/7. I have also found that a few of the so-called high efficiently PSU products make very little difference at all.

What I know for a fact is my electric rates keep going up and will continue to do so as will yours. Perhaps with enough info we can determine if there are one or two outstanding PSUs out there and which overall configurations run most efficiently as well.

I have noticed that the SMP client does use measurably less power but it would be fun and possibly very useful to compile this data and post it in sticky form and update from time to time.

Now, where did my wife hide my KilloWatt……………. ;)
 
ICE_9 said:
Watts: 122W
Amps: 1.6A
VA: 185
PF (Power Factor): .66
Total kWh used: 2.89
Total time: 23h 51m

Total cost per week to run this setup for a week($.087 per kWh): $1.77

Details:
Intel Pentium 4 2.4GHz (OC to 2.66GHz)
ECS PT800CE A
Geforce 2 MX video card
1 GB DDR ram
6 GB Maxtor HD
---------------------------------------------------------------------------------------------------------
Watts: 200W
Amps: 2.8A
VA: 340
PF (Power Factor): .59
Total kWh used: 7.17
Total time: 36h 27m

Total cost per week to run this setup for a week($.087 per kWh): $2.88

Details:
Intel Pentium D 3.0GHz ES
ECS mobo P4M800Pro V1.0A
on board vid
1 GB DDR ram
6.7 GB Maxtor HD
------------------------------------------------------------------------------------------------------
Watts: 46W (screen off) 59W (screen on)
Amps: .4A
VA: 47
PF (Power Factor): 1 - .99 (not bad)
Total kWh used: 1.66
Total time: 35h 45m

Total cost per week to run this setup for a week($.087 per kWh): $.68

Details:
Dell Inspirion 9300
Pentium M 1.73 GHz
1GB DDR2 RAM
Nvidia Go6800 Vid card
No USB devices connected.
-------------------------------------------------------------------------------------------------------
Watts: 232W
Amps: 2A
VA: 235
PF (Power Factor): .98
Total kWh used: 6.5
Total time: 27h 50m

Total cost per week to run this setup for a week($.087 per kWh): $3.42

Details:
Enhance 5150GH power supply
Intel P4 3.6GHz
2GB DDR2 RAM
MSI Radeon X800 Pro
DVD writer
Sound Blaster Audigy 2
Floppy drive
Usb devices: Bluetooth adapter for mouse, G15 keyboard, MS Sidewinder Voice.
* please note 4 hours of this test was during a run in BRD on WoW.
---------------------------------------------------------------------------------------------------------------
Watts: 162W
Amps: 1.38A
VA: 162
PF (Power Factor): .99
Total kWh used: 4.2
Total time: 26h 12m

Total cost per week to run this setup for a week($.087 per kWh): $2.34

Details:
Pentium D 830 ES
ECS P4M800Pro V1.0A mobo
Hard drive
Enhance 5140GH 80 Plus power supply (cost $63)
Running SMP client

Cost difference per week vs a Raidmax 350W power supply: $0.54 (cost for a 420W is $27.10)

Total time the Enhance needs to be in service to cover the additional cost of a efficient power supply: 66.5 Weeks in service.
-------------------------------------------------------------------------------------------------
Watts: 150W
Amps: 2.31A
VA: 278
PF (Power Factor): ..54
Total kWh used: 10.66
Total time: 70h 34m

Total cost per week to run this setup for a week($.087 per kWh): $2.21

Details:
Pentium D 805
ECS P4M800Pro V2.0 mobo with onboard video
Hard drive
Craptacular power supply
Running SMP client
--------------------------------------------------------------------------------------------------------
Watts: 144W
Amps: 1.17A
VA: 145
PF (Power Factor): .99
Total kWh used: 10.39
Total time: 72h 53m

Total cost per week to run this setup for a week($.087 per kWh): $2.08

Details:
Pentium D 805
ECS P4M800Pro v2.0 mobo with onboard video
Hard drive
Antec Earth Watts EA-500 power supply
Running SMP client

Cost difference per week vs a Crappy 350W power supply: $0.13 (cost for a 420W Raidmax power supply is $27.10)

Total time the Antec needs to be in service to cover the additional cost of a efficient power supply: 176 weeks in service
Click to expand...

Here are my findings. I think I need to redo the power figures. I am paying more for electric now. The higher efficiency doesn't always make a HUGE impact. But it does all me to put more systems on one circuit.
 
BillR said:
Although every KWH costs money my findings are folding adds very little to the overall cost of leaving the computer running 24/7. I have also found that a few of the so-called high efficiently PSU products make very little difference at all.
Click to expand...

Bill,

I now recall that I did check the power consumption with the Killawatt again after I put in the high efficiency power supplies (for the record, 1x Earthwatts 400w, 2 Earthwatt 380s), and power consumption did drop about 10-15%, depending on what each PSU was replacing. So I don't think the Earthwatt units are a bunch of hooey; I did write down those figures (but promptly lost them). In any case, any future boxen will be getting the high-efficiency PSUs.

Between retiring 3 XP2700 boxen, and tweaking one of my GPU client boxes (reduced core speed and voltage), I was able to save over 200 watts, even factoring in adding a second GPU-only boxen. So it was a win-win for me.
 
Opteron 165 @ 2.4 GHz (266*9)
2x 512MB (1GB) Ram @ 171.5 (DDR 343)
250GB Hitachi IDE 7.2K
250GB WD SATA 7.2K
9600 Pro
Ancient Enermax EG651P
19" Beast NEC FE991SB


Power Factor: 0.80
Idle Watts: 192
Idle VA: 239

Loaded Watts: 249 (180 w/o monitor)
Loaded VA: 311 (255 w/o monitor)

Running Windows SMP for about ~800 PPD
 
Couple of updates on what I have found. I have tried a few of my other P4 Northwood systems and all have about the exact same results as the first one I posted. Next one I have tested is the new Dell workstation.

Dell Precision 690
1KW Power Supply
Quad 15K SAS hard drives
ATI Radeon x1950gt 256MB
Dual - dual core 2.66ghz Xeon 5150
16x1GB FBDIMMS (uses a lot of power)

Idle - 2.51A - 301W
FAH SMP Client and FAH GPU Clients Running - 3.4A - 410W
 
You must log in or register to reply here.
Back
Top