Power efficency of gaming PCs is bad

[Dribble]

The thing is it's only using a lot of power if the person is playing a graphically intensive game, the rest of the time they all clock right down and are pretty efficient.

If they wanted to do a paper on ways to use a PC to uselessly waste the worlds resources they should have done it on bitcoin mining. A ponzi scheme that works off wasting power - the cost of the coin is literally tied in with how much power you had to waste to generate it.

 

[Madpacket]

Moving from a full ATX board to ITX and from a 650w PSU to a 380w dropped my idle power consumption in half.

This...

A lot of my friends who build gaming systems buy oversized power supplies which waste tremendous amounts of power over time . Considering these machines spend most of the time idling and on average 1-2 hours a day of gaming (probably less, depends on age, lifestyle etc) a change to a properly sized power supply can make a massive difference over the lifespan of the computer.

My HTCP (see below) idles around 40 watts and hits around 180-200 watts under heavy gaming loads because I used a properly sized power supply and this system is plenty fast for 1080P gaming and other tasks. This using a cheap Corsair 430W Bronze rated power supply.


Geforce 960, 16GB DDR3, MSI ITX motherboard, Xeon E3-1231V3 CPU, 2 SSD's, one 6TB and one 4TB HDD, and several fans all stuffed in a small Node 304 case.

 

[Madpacket]

If they wanted to do a paper on ways to use a PC to uselessly waste the worlds resources they should have done it on bitcoin mining. A ponzi scheme that works off wasting power - the cost of the coin is literally tied in with how much power you had to waste to generate it.

Really? People still think this way about Bitcoin? Wow..

 

[ShintaiDK]

My HTPC is currently based on an i5 661. It idles around the same as my gaming PC.

A NUC HTPC will reduce this with a factor of 10. Same factor I guess when it plays movies with half or unsupported codecs. Just in idle power the NUC will reduce the power consumption with almost 400Kw/h a year. Or something like 135$ in my case a year.

But again, a gated discrete GPU should reduce a modern gaming PCs idle power in ~half.

 

[.vodka]

Modern hardware has lots of sensors. Are these somewhat accurate? If so, my rig for example:

It doesn't look like it would consume that much. Even considering an old PSU that's 75% efficient at these loads, that would be what, around 50w being generous for the entire setup?

A new, more efficient PSU and a newer platform could probably cut that number in half. I'll have to find a way to measure what's going into the PSU (no off the shelf killawatt around here, and we use 240v) but yeah, the numbers in that paper seem somewhat inflated. The fact they used Sandy-E could explain why.

 

[TeknoBug]

My HTPC is currently based on an i5 661. It idles around the same as my gaming PC.

A NUC HTPC will reduce this with a factor of 10. Same factor I guess when it plays movies with half or unsupported codecs. Just in idle power the NUC will reduce the power consumption with almost 400Kw/h a year. Or something like 135$ in my case a year.

But again, a gated discrete GPU should reduce a modern gaming PCs idle power in ~half.

My NUC idles less than 2W and full load around 18W due to USB devices and SSD. It's a sweet system and I'd love to get a i7 5557U or i5 4570R.

 

[jpiniero]

The thing is it's only using a lot of power if the person is playing a graphically intensive game, the rest of the time they all clock right down and are pretty efficient.

Problem is that the typical PSU is inefficient at low loads. Going to something tiny would save a ton of power at idle. That wouldn't work with a power hungry dGPU though.

 

[poofyhairguy]

[ Most people would be better off with a cheap power efficient PC (even something like Skylake T which is hardly slow along with a power efficient mobo and DC power?) and then use the gaming PC for just games.

That is exactly what I tried to do a few months ago in my own life. My gaming PC became basically a fancy console in my livingroom only turned on for gaming and I built a Mini ITX modern Pentium dual core rig (3GHz, SSD, 8GB ram, etc.) for doing my "desktop tasks."

To my surprise I didn't like the decrease in desktop power going from my overclocked 2500k to an efficient modern Pentium dual core. I guess I multitask more than I thought. Ended up just hooking back up my old overclocked q9550 system after a while and leaving it running all day, which was probably the worse case scenario for energy use.

Now my plan is to upgrade my gaming rig so the 2500k can move to my desktop, and then to not overclock that desktop so when I leave the 2500k running all day at least it can be clocked down. At the end of the day my mediaservers and their combined 26 drives eat more power 24/7 than even my gaming rig could. Sorry Ms. Mother Nature, but I have decided to quit trying.

 

[piasabird]

Web browsing can be misconstrued. That could be included with watching kitty cat videos and Music on YouTube. Almost all websites now have video here and there. Now a lot of sites even on YouTube have HD video or may offer 3D or Ultra HD. I watch videos in HD and I have no reason to even use a video card. Some integrated computers can draw like 18 watts at idle and about 60 watching a DVD. To really get the power draw down the Motherboard selection might be a key component.

Using an intel 4330 with 4 Megs of Cache and 4600 Graphics wireless centrino card on a Gigabyte wifi Z??N-WIFI Motherboard attached to a HDTV. The HDTV is doing all the amplifying of the HD audio assuming it is allowing pass through.

I am considering just building the next computer with no hard drive, using SSD only.

Korean Videos with sub-titles:
http://www.tiveee.com/ep/scholar_of_the_night/?ep=1

 

[ShintaiDK]

Modern hardware has lots of sensors. Are these somewhat accurate? If so, my rig for example:

It doesn't look like it would consume that much. Even considering an old PSU that's 75% efficient at these loads, that would be what, around 50w being generous for the entire setup?

A new, more efficient PSU and a newer platform could probably cut that number in half. I'll have to find a way to measure what's going into the PSU (no off the shelf killawatt around here, and we use 240v) but yeah, the numbers in that paper seem somewhat inflated. The fact they used Sandy-E could explain why.

You only have a proper number for the CPU itself. Thats why a wattmeter is so important. An old PSU could easily be 20-50% efficient in the idle mode ranges.

 

[Cerb]

https://docs.google.com/viewer?a=v&...VlbmluZ3RoZWJlYXN0fGd4OjdhNTI3NzQ0MTJjYzY3MDE

See the paper. I don't necessarily agree with the paper, but it is kind of obvious that gaming PCs are a terrible waste of power if you are using it for anything other than gaming. Most people would be better off with a cheap power efficient PC (even something like Skylake T which is hardly slow along with a power efficient mobo and DC power?) and then use the gaming PC for just games. Then again, if you only use your $1000+ gaming PC for just games, isn't that basically a really expensive console?

If you did that, then you've got two PCs running. How's that an advantage? You're using as much extra power doing that as just wasting it through the idle GPU. The added crap will make up a lot of the difference (such as a KVM), which is going to be a single light bulb's worth anyway, unless the comparison is to a NUC or similar.

The best solution would be Optimus on the desktop, for the dGPU problem.

I use my integrated Xeon E3-1245V3 graphics because even in internet browsing, email, and stupid stuff a descrete graphics card is adding an additional 100w or so more.

Not with Maxwell, and an ad-blocker. I don't break 100W without either synthetic testing, compiling, or gaming, much less 100W just from the GPU. My old GTX 460? Er...it sort had idle and current hog modes only, despite being similar in TDP . W/o blocking ads, I can absolutely believe their web browsing numbers in the paper, though. How people manage like that, I do not know.

GPU-limited gaming tends to get me to around 300W total (80+ Platinum PSU, B85M Pro4, Xeon E3-1230V3, GTX 970, U2414H). I have little doubt that my motherboard is the weakest link in idle power consumption (it's a run of the mill 3+1 phase VRM setup, IIRC). Highest I recall was just over 320W, though usually being closer to 280W.

I think they were excessive with their Fig 14 improvements, in that an i5 and GTX 970/980, with a decent mobo and PSU, and any monitor normal people buy, would have gotten them 95% of the way to using a Pentium, without the minimum FPS sacrifices (which they do not document).

That said, I do agree with their conclusion, which is that non-techies don't have access to solid information on the subject of any given component's energy efficiency.

 

[zir_blazer]

When not running a game, most modern video cards only add 10-15w to the total system power consumption. Most modern components don't draw much power at idle. I imagine the big offenders at low-load are probably desktop motherboards and power supplies. Moving from a full ATX board to ITX and from a 650w PSU to a 380w dropped my idle power consumption in half.

15W is a lot for something that could be 0 if it were implemented properly. And if most people says that they computers Idle at around 50W, you can say that 15W is pretty much 1/3 of that.

 

[NTMBK]

If they want to replace my 130W Phenom II with a lovely power-sipping i7-5775C, that would be a-okay with me. PM me for shipping address, hippies.

 

[Madpacket]

People generally care more saving money so over saving mother earth. It's a selfish attitude but a realistic one. Most corporations work the same way. If you live in area with cheap electricity (10 cents or less) most of these issues don't even register.

Although I live in an area with relatively cheap electricity I still shoot for the most efficient and capable computer, and not to be responsible for green reasons (yes I'll admit it), I do it more for the reduction of heat and noise. Low powered computers also last longer, less wear and tear on the fans as they don't have to run at a high RPM, reduction of buildup of dust etc.

 

[SteveGrabowski]

Rather than regulations on SLI/CF, could we instead see a "cash for clunkers" program, for older, less energy-efficient PC components like GPUs? Free gov't-sponsored GPU trade-in program, everyone gets a free upgrade to 16nm Pascal?

hahaha

Seriously, no need for regulation here, if you can't get power consumption in check you can't get in mobile, and your company is dead in the water.

 

[monkeydelmagico]

I'd pit my sig rig against an XBox one for both idle and web neffing and win. Power efficiency of gaming PC's is only bad when viewed in a bubble of self righteous fart gas. The author can keep telling himself it doesn't stink but when the bubble bursts.....

 

[JimmiG]

When not running a game, most modern video cards only add 10-15w to the total system power consumption. Most modern components don't draw much power at idle. I imagine the big offenders at low-load are probably desktop motherboards and power supplies. Moving from a full ATX board to ITX and from a 650w PSU to a 380w dropped my idle power consumption in half.

Well a higher-wattage PSU will be more efficient, although using a very overpowered PSU, the system might not even draw enough to reach peak efficiency.

The thing with microchips is that pushing them to the limit causes their power usage to increase more than the performance does. For example, my work laptop has a Quad core Haswell at 2.4 / 3.4 GHz that draws 47W max. My desktop i4770K at 3.5 / 3.9 GHz draws 84W. That's a ~15% higher max turbo but nearly twice the amount of power used. Overclocking it to 4.5 GHz probably increases that to close to 200W, for another 15% of performance.

 

[know of fence]

The curious thing about PC gaming is that in many ways it developed like mobile and free-to-play games. There is the mainstream that pays little or doesn't buy dedicated cards at all, but there is a small percentage of "whales" that are now targeted with 500 and thousand dollar CPUs and graphics cards.
Moreover heat-pipe cooling allowed for 250W TDP cooling solutions, and AMD is even trying to push 300W with liquid cooling now.

Thar she blows!

 

[ClockHound]

Just for fair and balanced reporting...perhaps Google could state the massive power used by their consumer data mining farms. Or do those all run on a handful of Chromebooks now?

 

[Cerb]

Just for fair and balanced reporting...perhaps Google could state the massive power used by their consumer data mining farms. Or do those all run on a handful of Chromebooks now?

I'm sure they are heavily optimized for energy efficiency. Increases in W/m^3 translates directly into added ongoing costs, but there will be a point where decreasing it would mean losing too much in terms of data processed over time. Power needed for a given space is a big deal, and improving it by just a few percent per datum processed across a data center might well save them tens of thousands per month per location.

 

[SPBHM]

you can build a gaming PC with an i3 and 750 ti which will give you lower power usage than a PS4 for the same graphics,

 

[MongGrel]

Well a higher-wattage PSU will be more efficient, although using a very overpowered PSU, the system might not even draw enough to reach peak efficiency.

The thing with microchips is that pushing them to the limit causes their power usage to increase more than the performance does. For example, my work laptop has a Quad core Haswell at 2.4 / 3.4 GHz that draws 47W max. My desktop i4770K at 3.5 / 3.9 GHz draws 84W. That's a ~15% higher max turbo but nearly twice the amount of power used. Overclocking it to 4.5 GHz probably increases that to close to 200W, for another 15% of performance.

Going way over wattage on a PSU isn't efficient.

Even though I still use an Antec CP-850 on the main rig that isn't really needed these days.

http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story&reid=142

Had more GPU's in here at one time.

 

[Sable]

Power efficency of gaming PCs is bad[/QUOTE] im shocked. no rly. O_o

 

[Yuriman]

Well a higher-wattage PSU will be more efficient, although using a very overpowered PSU, the system might not even draw enough to reach peak efficiency.

The thing with microchips is that pushing them to the limit causes their power usage to increase more than the performance does. For example, my work laptop has a Quad core Haswell at 2.4 / 3.4 GHz that draws 47W max. My desktop i4770K at 3.5 / 3.9 GHz draws 84W. That's a ~15% higher max turbo but nearly twice the amount of power used. Overclocking it to 4.5 GHz probably increases that to close to 200W, for another 15% of performance.

Those are TDPs, not actual power consumption.

You might find this chart interesting from my thread on power consumption:

If we assume ~5w idle draw, under an AVX load and accounting for PSU efficiency @ ~80%, an Ivy Bridge desktop chip draws ~12w at 1600MHz, ~21w at 2600MHz, and ~36w at its maximum non-turbo frequency of 3400MHz. Going up another 600MHz increases power consumption by almost 50%, which makes it pretty clear why the stock frequency is what it is.

 

[Vesku]

I think we all know that gaming PCs specially are hummers in that regard.

I am surprised we dont have setups, when the vast majority got IGPs. So the discrete GPU is turned off for example under any non gaming.

My system idles (browsing etc usage) around 40-50W or so. And I guess half of that is from the GPU alone.

Compatibility issues trying to support the full variety of PC configurations. Lucid Virtu MVP doesn't work for me (Windows 7), some sort of resource conflict that wasn't solved by swapping the dGPU's PCIe slot nor removing and re-adding the Intel IGP. Perhaps at some point the display port or PCIe standard will work out a reliable way to handle this. I'm guessing Windows 10 handles multiple GPUs better than previous versions.