Power SupplyMatchup: PS3 Vs. the Quadfather

[robwright]

Gaming takes a lot of energy, but power supply units are often overlooked as an integral part of the hardware scheme. TwitchGuru compares the PSU of the PlayStation 3 to the PSU of a top of the line quad core gaming rig, to see how much power each system uses during different tasks.

 

[joex444]

We ran a Windows power management test on the system, which set off the Quadfather's heat alarms as one of the CPU's temperatures rose to 160 degrees Fahrenheit; the second CPU soon crept over 100 degrees. However, it turns out the Quadfather's massive exhaust ports were positioned too close to a nearby wall, causing hot air to be blown back into the system. With proper ventilation, the Quadfather avoided overheating during testing and gameplay.

That should never have been mentioned. It is YOUR fault for not knowing how walls work, you rectified the situation and the Quadfather worked as you wanted it to.

 

[HotFoot]

I think it's good to read about the little screw-ups that reviewers do. It helps the rest of us avoid the same situation. It's not like they held it against the QuadFather.

What gets me about this article is that they are using the QuadFather in the first place. They admit that this computer is well known for having horrible power consumption, but really, the QuadFather? This hunk of junk is not a gaming platform. It's a workstation in an enthusiast's computer's clothing. I'm quite certain that computers using half the power can provide superior gaming performance.

Admittedly, AMD's Quad FX platform doesn't represent the company's finest hour; it has been criticized for its sizeable increase in power requirements compared to the Intel Core 2 duo. Nevertheless, our assessment demonstrates that the PS3's power consumption and PSU are impressive when stacked against a PC gaming rig of similar strength.

The first sentence there starts of well. Then the second sentence is just crap. The Quad's gaming performance leaves the PS3 in the dust. If you want a head-to-head comparison, why not team a 65W E6600 with a single 8800gtx. That combination should give you equivalent gaming performance. Folding@Home and other HPC projects aren't gaming. If you want energy efficient Folding, then by all means the PS3 is the winner.

I was also thrown off by the power requirements always including the monitor. I understand that comparing the gaming system power consumption means the entire system, including video and sound. I think it's a good idea to talk about it, but what gets confusing in the article is how the total power consumption is mentioned in close quarters with the PSU rating. The PSU isn't powering the monitor. It would be good to note that the QuadFather consumed about 780W, which is a very significant margin below the 1000W PSU rating. I would, though, be more concerned with how many amps the 12V rails are providing w.r.t. their rating.

I think the really useful comparison here is between the PS3 and the Xbox: console to console. Otherwise, give us power consumption for a computer that's at least in the same ballpark in terms of gaming (not floating point) power.

 

[bkiserx7]

Anyone got any links to the PS3 folding?

 

[HotFoot]

You can get an approximation of what the contribution of the PS3 is to Folding@Home at the [urlhttp://fah-web.stanford.edu/cgi-bin/main.py?qtype=osstats]Client Statistics by OS[/url] page. Since the PS3 has come on the scene, it seems that the total TFLOPS of the project has doubled.

It should be noted, though that TFLOPS is not a hard-and-fast measurement. In this case, it's an approximation based on the requirements of the software, not the actual operations that the CPU does. For example, some software operations must be broken down into two steps to execute (so 2 FLOPS on the CPU might mean only 1 FLOPS in software).

 

[Athalus_nubie]

This article is an absolute shocker. Why compare the quadfather to a ps3?
Why include the wattage the monitor draws, seeing as its not drawing that from the PSU in the first place?
What did this article set out to prove, or say? it was just a few pages of "the ps3 isnt as crap as some people have said"

 

[kuyaglen]

Nice. I would of liked more pictures though and maybe wattage used (by the PS3) when its running linux, downloaded demo's and older games (assuming that older games would tax the gpu as much and use less power).

 

[HotFoot]

I'm not too sure what the article actually set out to discover. I think comparing the power consumption of these two systems is a little odd. Why not compare the pricing?

I think there is a value to mention the power required by the monitor. If you're very conciensious about the power you're consuming, then you want to be looking at the gaming experience/total system power. What that has to do with the PSU... well, there's a degree of separation there.

Maybe it all started with some historical perspective about how systems are starting to consume massive amounts of power, both on the console and extreme gaming (or in this case I still assert workstation) sides. It would make just as much sense for me to compare the consumption of the PS3 with my media centre PC, which I'd estimate to use about 100W idle, 150-160W full load. I'm really starting to think I should get myself a watt meter and take some measurements. It would be especially good to know considering air-conditioning season is starting... also for better estimating the cost of running folding@home 24/7 on two PCs and a laptop.

In general, any comparison between consoles and computers just takes too much of a stretch of the imagination. I recall a few times when people have come onto the forums trying to settle an argument about which type of platform is the best for games.

 

[robwright]

That should never have been mentioned. It is YOUR fault for not knowing how walls work, you rectified the situation and the Quadfather worked as you wanted it to.

We mentioned it because, even though we thought we had enough room for the system to breath, the Quadfather overheated within a matter of seconds after turning the system on for the first time and tasked with running a simple power management test. Don't get me wrong -- it IS our fault for having the Quadfather positioned wrongly for ventilation. But we were surprised that it overheated so quickly. And as HotFoot pointed it, we throw these kinds of oversights and screw ups in because, hey, we don't want readers to go through the same thing.

Here's the latest on the folding@home tests:
http://www.tgdaily.com/content/view/31778/98/

 

[robwright]

To the questions about why we compared the two, here are some answers. First, why not? We had a PS3 in the office and we just got the Quadfather for testing. We decided it would be an interesting comparison to take the newest console with a 3.2 Ghz processor and stack it up against one of the newest gaming rigs with a 3.0 Ghz processor.

Comparisons between PC gaming and console gaming always seem to get people talking. Yes, it's a bit like apples and oranges, as you can never really make a straight comparison. But both the PS3 and the Quadfather are designed to deliver the best performance possible in their respective areas of gaming, so a matchup likes this illustrates some of the differences between the hardware involved.

 

[r0x0r]

Comparisons between PC gaming and console gaming always seem to get people talking. Yes, it's a bit like apples and oranges, as you can never really make a straight comparison. But both the PS3 and the Quadfather are designed to deliver the best performance possible in their respective areas of gaming, so a matchup likes this illustrates some of the differences between the hardware involved.

What would have been the fairest thing to do is to wait for a game to come out on both PS3 and PC and bench them both (power and fps) at the same resolution (i.e. 1920x1080) and settings. Then you divide the power draw in wattage by the average fps (i.e. 750W / 60fps = 12.5W/frame). The lowest watt per frame is therefore the most efficient.

 

[stemnin]

I want to.. but I just can't insult the quadfather... it scares me.

 

[nukemaster]

just for fun you should put together a Q6600(how the prices have fallen!!) system with 1 8800GTX and check out the power. I am just saying this as I am sure you got a ton of parts to play with

 

[TeraMedia]

If the PSU for the Quadfather includes components for 3.3V, 5V, and 12V rails. It can't put out all 1000 W towards any one line or voltage level. Also, there is the energy efficiency to contend with; (power out) = (power in) - (heat and other energy drains). So if it's using 860 W at the wall (or is it 760 W? Your inclusion / exclusion of the monitor made measurements unnecessarily confusing), then it's probably using a max of 0.85 * 860 or 730 W internally. That 730 W implies a hefty margin of safety compared to the 1 kW rating. But you have to consider that there are limits on individual rails so that supply may have been necessitated by the dual CPUs, or the dual 8800 GTXs, even though the supplies for 3.3 V and 5V have plenty of room to spare.

So: the individual line output requirements inside the Qfather need to be considered when evaluating the choice of PSU, and margin of safety/error. I don't think your article talks about that enough.

Also, the PS3 will likely have similar limits on individual lines, but its power supply was likely pieced together from parts of commercial PSU components, according to the needs of the HW. Given that, I would expect it to use a smaller PSU than the rated 380 W (which represents ~450 W at the wall) - as you demonstrated in your article. The PSU is far larger than necessary, which means that the PSU is wasting more power than it needs to at all times, and particularly at idle. That's what tells me that the PSU for the PS3 isn't very energy-conscious. Not the fact that it uses less power than a Qfather.

 

[HDDFreak]

Why include the wattage the monitor draws, seeing as its not drawing that from the PSU in the first place?

Umm, I think you are missing the point here. How can you test the PS3 during gameplay without a monitor :roll: ? Granted, they could have plugged the monitor into another powerpoint (if they had one), but what would it have accomplished.

I thought it was an interesting article. It is certainly different to what is normally published here.

 

[Shin-san]

Tomshardware seriously should take video of some of their massive rigs and run video capping and then upload it to gametrailers.com

I would love to see how Oblivion looks with their 8800-SLI rig at 1080p