How much power does your PC use?

[BUCK NASTY]

I take it that's with the opterons running overclocked? 780w seems a little high for stock but I could be wrong.

Yes, there is a 13% overclock. Opty 6180se stock is 2.5ghz and running 2.83ghz on stock voltage.

 

[Norton]

I take it that's with the opterons running overclocked? 780w seems a little high for stock but I could be wrong.

My 4P runs 6168's at stock and draws about 570w at the wall. I'm using an old 1kw Silverstone psu atm.

 

[RejZoR]

Yikes! I have the same cpu as you and a 290X and my system idles at around 78w. How much of a difference a few hard drives make.

Haha, 78W idle. I wish. Mine is at 250W idle. 4GHz is a bitch.

 

[Aquinus]

Yikes! I have the same cpu as you and a 290X and my system idles at around 78w. How much of a difference a few hard drives make.

Yup, the 5 HDDs are probably about 50-60 watts. Also remember that I have 3 displays, not 1. AMD's multi-monitor idle usage is far from stellar. I bet with just one display my consumption would drop 50 watts, just 1 HDD instead of 5 is dropping another 50 watts. Drop the second 6870 in low power mode and you have your 80 watt idle more or less.

It's super important to remember that my high idle usage is a result of everything I have in and attached to my tower, not the CPU. The CPU actually sips power at idle and even at stock (which it's at now, because I've been too lazy to OC it again.)

 

[rtwjunkie]

My server listed in sig draws 113w, jumping to 130 if it is streaming movies. I've got 8 HDD in it and an SSD, as well as an HDD external, all coming through the UPS.

I will have to test my main rig for actual numbers using 3dMarl like is being done here, but I can say, anecdotally, that last night for instance, playing Rome II, the UPs was showing only 386w, which I thought was impressively low!

 

[Nokiron]

Remade the test for reference against my old ones.

Old:

Nokiron | Corsair AX1200i | i7 3970X@4.1Ghz 1.175V - Gigabyte GTX 780 Ti 1262/7500 | 121W idle | 226W | P14705 415W | 35.43 points per watt

New:
Nokiron | Corsair AX1200i | i7 4960X@4.5Ghz 1.285V - Inno3D GTX 980 | 135W idle | 241W | P16727 350W | 47.79 points per watt

Damn, that is a pretty significant increase in points per watt.

 

[Mussels]

Remade the test for reference against my old ones.

Old:


New:
Nokiron | Corsair AX1200i | i7 4960X@4.5Ghz 1.285V - Inno3D GTX 980 | 135W idle | 241W | P16727 350W | 47.79 points per watt

Damn, that is a pretty significant increase in points per watt.

added it to the list - and yeah, stuff like that is exactly why i do these tests. some people would be all "oh no its using more power at max, i've wasted my life" - but if you're getting double the performance, you can always use Vsync or an FPS cap to get a more efficient system out of it anyway.

 

[NateDawg]

I've just replaced the GTX 660 OEM from my previous entry with a R9 290. Fiddling around with clockspeeds and voltage today, then I'll tally up the results in the next couple days.

Yikes, I completely forgot I said I'd do that once I got busy with the holidays. Anyway, here it is.

NateDawg | SeaSonic SSP-450RT | i5 3350p @ 3.5GHz + Asus R9 290 | 53w idle | 85.5w | P11826 279W | 42.39 points per watt

Changes since last test:
+200mhz CPU clockspeed, small voltage bump
replaced 660oem with an Asus R9 290.

Other Notes:
-290 is at stock speeds (Asus set it higher than reference - 1000mhz core / 1260mhz memory)
-CPU & GPU are both undervolted.
-290 doesn't like running at 144hz - at that refresh rate the vram doesn't idle and stays at 1260mhz, resulting in roughly 92w total system idle and low-mid 40s temperature. Dropping the refresh rate to anything else shaves about 40 watts off as the vram idles at 150mhz and temperature drops to mid 30s. I'm using 120hz due to this behavior. Not sure if all GCN cards do this, or if it is 290/290x specific.

Old test for comparison

NateDawg | SeaSonic SSP-450RT | i5 3350p @ 3.3GHz + 660 OEM | 48w idle | 74.4w | P6097 141W | 43.24 points per watt | Daily use

Additional GPU undervolting/clocking info:
Had run some 3dmark11 benches at a couple other settings and found the results interesting, maybe someone else will too. I reduced the clock speeds and maintained the core-memory speed ratio of default clocks. A voltage offset of -69 was set in Afterburner for all my tests.

Test 1 - 1000mhz core, 1260mhz memory
P11826 279w | 42.39 points per watt

Test 2 - 900mhz core, 1134mhz memory
P11133 234w | 47.58 points per watt

Test 3 - 800mhz core, 1008mhz memory
P10308 210w | 49.09 points per watt

Some rather noticeable gains in efficiency if you ask me.

 

[ChristTheGreat]

Add my result
ChristTheGreat | OCZ Z-850W Gold | i7 4770k 4.3ghz + R9 290 1100/1400 | 95w idle | 205w | P14343 386W | 37.15 points per watt | 7 fan + pump, using dual-screen

 

[kn00tcn]

i recently got a killawatt, can finally get some measurements

FP241VW - 48 minimum brightness, 87 maximum brightness, ~60 with my preferred settings (58.8 all black image, 60.9 all white image), yes this is CCFL instead of LED

PS3 - ~104 minimum with nothing happening (photo viewer), ~110/115 XMB, ~125 pshome or 'low load' with 30fps capped etc, ~138 max load (tearing in GT6 aka max gpu usage), this is a CECHL, the last 80gb fat (65nm)

have not tested laptop or desktop yet

i'm curious what my newly discovered laptop OC will be (525 to 725mhz gtx570m)

too bad the 4870x2 died, can imagine hundreds of watts.... but that desktop is now gtx660, i guess it would be interesting to find out how much my q9550 3.6ghz OC uses compared to .... well it's never ran stock, & my manual voltages are almost identical to the auto's overvolting

 

[Outback Bronze]

G'day Mussels got a new one for you mate. I got off my ass to test this system. Over 50% efficiency. I could definitely tweak that to get it higher as well!

Outback Bronze | Corsair AX1200 | i7 5960x @ 3.3Ghz + GTX 980 | 114w | 153w | P18092 336w | 53.845 | Stock Settings.

This power supply is also running 1 pump, 4 UV lights, 2 Storage Hard Drives and 5x140mm fans.

 

[agent00skid]

Add my result:
agent00skid | Laptop brick | Phenom II N830 + Mobility HD 5650 | 21w | 55w | P1014 59W | 17.19 |

And for fun, 3 threads of P95 and Furmark at high priority managed to push out 87w.

 

[Sleepless]

Here is my result. I ran my 4690K undervolted and it clocks down on idle, the 770 was at stock clocks even under boost but still gets up to 1.2V.

Sleepless | EVGA 750 G2 | i5 4690K @ 3.9GHz + GTX770 | 43w | 88w | P10230 299W | 34.21 | 4690K @ 1.024V under load, GTX770 @ 1163MHz

 

[Frick]

Heeeeyy I've finally done it. The current was measured with a Fluke 43B with a clamp meter and multiplied with 230 (which actually is what I'm getting)

Frick | Cooler Master G550M | C2D E8400 @ 3 Ghz + ASUS GTX 760 OC model | 78w idle | 94w | P4561 228W (max reading) | 19,87 points per watt (max reading) | 0.99 x 230

 

[Mussels]

i need to throw in my non-gaming ultrabook into this, it uses less power watching 1080p videos than my desktop does idling.

 

[Frick]

Are we still doing this? Because I finally have one of those wall meters.

Frick | Cooler Master G550M | Pentium G3220 + ASUS GTX 760 OC model | 56W idle | 77W | P5876 238W | 24.7 points per watt | n/a

 

[Mussels]

god, it'd take me hours? days? to re-test all my machines... i'm upto about 8 gaming systems right now for some unknown reason (i have a hoarding problem. that Q6600 needs love)

still happily updating the OP, as its an interesting reference, even if we found flaws in the testing.

 

[Jetster]

god, it'd take me hours? days? to re-test all my machines... i'm upto about 8 gaming systems right now for some unknown reason (i have a hoarding problem. that Q6600 needs love)

still happily updating the OP, as its an interesting reference, even if we found flaws in the testing.

And I feel bad about having 4 systems. I live alone

 

[Mussels]

And I feel bad about having 4 systems. I live alone

i5 2500k + 290
i7 2600 + 7970
i5 2500 + 6870 1GB (waifus PC)
i5 2400 + 5870 1GB
i5 2400 + GTX 280
athlon II x4 + 6570 2GB
athlon II x3 + 6570 2GB
phenom II x2 + 550 ti
Q6600 was a joke. i have one, but its not in a system. just... being itself. (i spent $600 on my first one, and got this free. its for ironies sake)


I... tried to downsize? i don't know how this happened.

 

[Kursah]

I have my server hooked up to a UPS with my networking hardware...and itself consumes anywhere between 90-130W depending on load conditions, usually settling around 94W in most situations. The networking hardware (cable modem, ERL, 8-port sw) consumes approx 12W. My old Dell LCD monitor consumes around 40W when it is powered on.

I saved power moving the core OS to SSD and setting the low cache multi to 8X, brought the idle down by almost 15W. Idle/low load (VM load) is around 90-115W. I run anywhere between 4-6VM's, MineOS Turnkey VM hosting 4 Minecraft servers, a 2012R2 DC, 2012R2 File/Print/Plex/Teamspeak server, Ubuntu 15.x test VM, 8.1 VM for deployment/domain testing, a PFSense VM for OpenVPN and network testing...all of those are running all the time. I have my server setup so I never need to login for anything to run on any of the 2012R2 servers...meaning Plex and Teamspeak run as services which is really nice.

Specs for my home server are in system specs...the HDD setup though has more than I could list...

Right now I'm running a storage pool with parity (2012R2 storage spaces) for backup (testing), so I have 2X 1.5TB and 1X 2TB dedicated to that. 4X 2TB in RAID 5 on the Perc 6i, 120GB SSD for OS. The RAID is in a cage that also has an extra 120mm fan, bringing 3x120mm intake, 1x120mm exhuast and top-mounted PSU exhausting as well.

Overall I'm pretty happy with this system...much quieter than any server I've worked with or deployed...granted its just a gaming rig sans graphics card...so when the day comes where I can afford the server I want, that's what this machine will turn into!

I plan to place my gaming rig behind a UPS eventually...we'll see how my tax refund looks...if I don't get a wall meter between then and now, I'll post what my gaming rig does...I actually want to test it stock, undervolted, and overclocked.

 

[Nokiron]

Why not update this for fun its really interesting.

Nokiron | Chieftec SFX-500GD-C | i7 6700K + R9 Nano | 46w idle | 90w | P15698 240W | 65.40 points per watt | GPU+CPU undervolted.

Nokiron | AX1200i | i7 4960X @ 4.5Ghz 1.25V + EVGA 980 Ti Classified | 181w idle | 265w | P21386 483W | 44.27 points per watt | Including full loop, 14-fans, 8 SSDs etc

The Nano-build must be some kind of record, Im really impressed with that.

 

[manofthem]

Spoiler

@Mussels I know it's been a while since this thread has seen action, but would you be interested in getting it going again to add some of the newer hardware to your table? With newer Intel chips and Ryzen out, along with newer gpus, it would be cool to see some of those numbers up there too.

 

[Mussels]

If people come up with the numbers i have no issue updating the thread - its just going to be slow going as i work 6 days a week now.

I have entirely new systems myself, that could be entered in here.

If we want to do some kind of overhaul (new programs for testing?) i'm all ears for suggestions.

 

[manofthem]

If people come up with the numbers i have no issue updating the thread - its just going to be slow going as i work 6 days a week now.

I have entirely new systems myself, that could be entered in here.

If we want to do some kind of overhaul (new programs for testing?) i'm all ears for suggestions.

Sweet, that's awesome! I'll update as soon as I get a new watt meter with my newer Ryzen system and a few other changes I've made.

As far as programs go, I have no real quirks. Seems people go with Cinebench for cpu? Mining is big but I don't know what people use to bench that.

 

[Mussels]

cinebench is a good idea for CPU testing imo, as it has single and multi threaded testing options - but since its openGL test isnt too useful, we'd need a GPU test as well.

Ideally i want to stick away from paid programs and something easy to use, maybe unigine heaven 4.0?