Apple's 24" iMac: There's more to High End than Performance

TESTING

The 24" iMac was tested twice: Once in OS X and once in Windows using our standard test tools. Unlike our test of the 17" iMac, these tests did not reveal any thermal or efficiency differences between the two operating systems, although the more robust power saving features in OS X do more for saving power than the default options in Windows.

The following software was used during testing:

OS X

• A custom-compiled version of CPUBurn
• Temperature Monitor 3.9

Windows

• CPUBurn
• Intel Thermal Analysis Tool v2.05
• ATI Tool v0.25 Beta 14

Power consumption was measured with a Seasonic Power Angel, and noise was measured with our usual B&K 2203 SLM.

Ambient conditions at the time of testing were 18 dBA and 21°C.

The 93W that the 24" iMac ate up at idle was a far cry from the 46W that the 17" model used when we tested it. A whopping two thirds of this power is consumed by the LCD screen, which is significantly brighter (400 CD/m vs. 250 CD/m) and more contrasty (700:1 vs. 500:1) as well as being larger than the 17" model. With the LCD screen in standby, the remaining components required just 31W — less than any desktop system we've built for ourselves, and, significantly, approximately the same amount required by the 17" iMac, despite the differences in hardware.

Things got better when the LCD dropped into low power mode by reducing the brightness to the lowest level possible. The amount of power consumed by the LCD dropped by almost half to about 34W, now about 50% of the total system power. However, the automatic use of this function sometimes got quite irritating, since the low power mode sometimes kicked in while reading a large block of text on a web page and the mouse would have to be moved around to boost the brightness back up. Given how vibrant the display is by default, the best solution is brightness down all the time.

If the system was left alone for much longer — about 15 minutes — the system would shut itself down into standby mode, where the total power consumption dropped to just three watts. Our experience has generally been that standby mode is more stable and reliable in OS X than in Windows, but that doesn't get around the fact that its not 100% reliable at detecting idle behavior. We had to disable the standby feature during testing because the system kept shutting down while we were running CPUBurn.

With CPUBurn going, power consumption jumped up about 30W — pretty much what we expected of the new Core 2 Duo chips, but almost twice as required by the Core Duo-based iMac that we saw. The processor quickly heated up to a toasty 66°C at full load. That's hot but not too hot, and we were pleased to note that the noise level never increased — it stayed at a barely noticeable 20 dBA@1m no matter how we stressed the system.

The noise character was the same unintrusive background hum that the 17" iMac exhibited. From memory, it seemed very slightly more noticeable, but we weren't able to measure a difference. Perhaps the difference is the hard drive; the 24" model used a Seagate instead of a Maxtor. In any case, then noise was quiet, easy to ignore, and didn't change during use. That's exactly what we expect of a quiet computer, and, despite the additional heat from the LCD screen, the 24" iMac delivered.

One thing we noticed was that the system seemed to be a bit louder when placed against a wall, presumably because the wall reflected the noise coming off the back of the system. This was confirmed with a measurement: The system measured 21 dBA@1m when moved within twenty centimeters of the back wall in the lab.

The noisiest the iMac ever got was when the optical drive was in use. Burning at 24x was quite noisy, but the noise character was surprisingly benign: Mostly the whoosh of airflow with a little bit of whine underneath. It sounded decidedly less harsh than the conventional drives that we are used to, most likely because the plastic casing of the iMac provides a bit of damping. Unfortunately, it did a poor job of damping the thunderous seek noise, which actually sounded more intrusive at 24 dBA@1m than the smooth burning at 30 dBA@1m.

The hard drive in our sample was a Seagate, so we were expecting noisy seeks, but the seek noise was surprisingly muted — it sounded less sharp than Seagate's drives sound in free air. The 1~2 dBA@1m rise above the baseline noise accurately captures what we heard: Seeks were clearly audible, but they weren't especially loud.

Thermal and acoustic performance in Windows was exactly the same as in OS X, so it merits no more comment. However, running in Windows did allow us to try one additional test with Intel's (technically unreleased) Thermal Analysis Tool. In addition to being able to monitor core temperature and throttling on Core 2 Duo chips, this tool does a better job of stressing the CPU than any other utility that we know of.

With the help of this admittedly unrealistic test, we were able to push power consumption up another 13W. The CPU temperature climbed to a toasty 75°C at this level, but even at this level the noise level never increased. We left the system burning in this way for over an hour, but the system was perfectly stable and the processor did not throttle. We can only conclude that, hot as it was, it never reached unsafe levels.

One thing that running in Windows did let us do was to monitor the GPU temperature. Surprisingly, the GPU seemed more affected by what the CPU was doing nearby than any graphics load we could come up with. Hence, the 20°C difference between idle and full CPU load, but the 0°C difference when we threw ATI Tool's artifact detector into the mix. We can only conclude that the mobile 7300GT chip doesn't produce enough heat to worry about. In any case, the GPU temperature was always below the CPU temperature, so we have no worries about its longevity in the iMac.

MP3 SOUND RECORDINGS

Each of these recording have 10 seconds of silence to let you hear the ambient sound of the room, followed by 10 seconds of the product's noise.

• Apple 24" iMac at any load, 20 dBA@1m: One meter, One foot (30cm)

• Shuttle X100 — Idle (Fan off): 23 dBA@1m: One Meter, One Foot
• Shuttle X100 — Idle / Load (Fan on): 25 dBA@1m: One Meter, One Foot

CONCLUSIONS

We liked the first iMacs we saw, and very little has changed with the 24" model so we like it as well. It remains one of the quietest off-the-shelf systems it is possible to buy and, with a built-in 24" screen, it offers fairly good value-for-money. A similarly configured Dell system that we looked at was only about US$200 cheaper, and it lacked the integration and style that are Apple's hallmarks.

It is also fairly efficient, although the large LCD prevented it from being as frugal as the 17" model. In fact, the LCD consumed the lion's share of the power — well over half when the system was idle. We're not sure how this falls in with other large LCD screens on the market, but it certainly consumed more than the 17~19" screens that we've measured, which typically require 25~35W. Discounting the LCD, the system was on par with the 17" iMac at idle, and only slightly more power hungry under load.

Is the higher power consumption a fair price to pay for the larger screen? Well, that's up to you, but we can certainly think of worse ways to consume power. The larger screen is potentially much more useful than, say, a faster processor or a high end motherboard. That's not to say that high performance components don't have their place, but most users — especially those who would buy an iMac — will probably see more benefit from a bigger screen than a higher clocked processor.

Besides, the iMac is far from a slouch. Intel's Core 2 Duo provides oodles of power, and the iMac was as speedy as any system we've used. Performance purists will no doubt complain that performance is bottlenecked by the slow FSB that goes with the mobile processor, but we noticed no performance hit. We just lament the fact that Apple is charging extra for a faster processor that won't be used.

All in all, the iMac is a high end product that the masses will understand. It's luxurious status comes from its image and integration, not performance. Thankfully, low noise is part of that image, and we like it for that all the more.

Much thanks to Apple Canada for supplying the 24" iMac sample for us to review.

*

SPCR Articles of Related Interest

Apple iMac w/Intel Core Duo: A User's Review
Fanless Ultra Powerhouse PC by EndPCNoise
17" iMac — The Official SPCR Review
Puget Delivers a Quiet Core Duo PC

* * *

Discuss this article in the Silent PC Review Forums.