Tests show iMac hard disk performance can vary by 100%

by Adam Banks on May 11, 2011

Apple has confirmed that it reserves the right to change the supplier and specification of the hard drives built into the latest iMac models, announced last week. In MacUser’s tests, undocumented variations between the drives fitted in different models resulted in transfer speeds ranging from 76 to 164 megabytes per second (MB/sec), making the quickest units twice as fast as the slowest.

The performance gap is invisible in Apple’s published specifications and is not guaranteed to be preserved in future production runs, meaning buyers can’t be sure what results they’ll see from the hardware they receive.

The 21.5in iMac supplied by Apple for testing, corresponding to the £999 configuration available from the Apple Store (2.5GHz Core i5), was fitted with a 500GB WD Caviar Blue drive (WD5000AAKS). In MacUser’s tests, this delivered an average transfer rate of 76MB/sec, peaking at 85MB/sec.

The supplied 27in model, as sold at £1399 (2.7GHz Core i5), contained a WD Caviar Black (WD1001FALS) 1TB drive which performed ahead of the manufacturer’s speed claims in our tests, achieving an average transfer rate of 164MB/sec and peaking at 173MB/sec. No details of the drives used are given in Apple’s published specs, which list only the same rotational speed of 7200rpm for all models.

MacUser researched additional iMac units and consistently found the same two drives used in the respective models. Among other configurations, we found the same 1TB WD Caviar Black unit in the 21.5in 2.7GHz Core i5 iMac, where it achieved average and maximum write speeds of 157MB/sec and 167MB/sec.

However, we found an entirely different drive in several 3.1GHz Core i5 models and in one example of the top-end 3.4GHz Core i7. This 1TB Seagate Barracuda 7200.12 unit (ST31000528AS) claims a sustained maximum transfer rate of 125MB/sec, but our tests showed average and maximum random write speeds of 101MB/sec and 111MB/sec: about one-third faster than the Caviar Blue, but one-third slower than the Caviar Black – which we also found in another 3.4GHz Core i7 iMac, nominally of the same specification.

Various factors account for differences in hard disk performance, including the amount of cache fitted (the WDC Blue drive had 16MB, while the faster WDC Black and Seagate had 32MB) and the number of platters used.

In disk-heavy tasks, these substantial differences will noticeably affect overall performance. In general use, paging apps and data in and out of RAM – a common background process for Macs with several heavyweight apps in use at once, especially with relatively small amounts of memory, such as the 4GB fitted by Apple as standard – will also cause more delays if the hard disk is slower.

Apple’s choice of drives doesn’t appear to correspond to the performance level of the system, so higher-priced models may have a slower drive. The variation between our two Core i7 models indicated that orders for the same specification could be delivered with different drives.

A spokesperson for Apple told MacUser: “As always we do not specify the HDD vendors, which allows us to get the best deal for the customer.” Customers may feel that they would prefer to know what deal they were getting before handing over their cash.

For more breaking news and reviews, subscribe to MacUser magazine. We'll give you three issues for £1
  • J400uk

    Seems to be a recurring theme. We have a room full of 27in iMacs at uni, they were all purchased the same time as far as I know and have the same specs, but they don’t use the same hard drives! Seems to be a mixture of WDs and Seagates of the same capacity. No performance benchmarks but one lot is definetely a lot noiser than the other.

  • jdickey1

    Um, yes. We do have a problem here.

    Since the iMac is the only desktop Mac for most people, (since most people can’t/would prefer not to fork out for the Mac Pro,) and since “getting the best deal for the customer” does not include passing the savings on to the customers who “got the better deals,” what we have here seems more than a bit like “good, old-fashioned” bait-and-switch on the part of Apple. I would seriously doubt that many of the demonstrator units in stores have the slower drives.

    As a long-time fan, customer and shareholder, this does not make me happy. Yes, the outsides of iMacs are pretty. Yes, the specs, so far as they go, are appealing. Yes, the software at many levels is legendary for its power and ease of use.

    But it’s still (largely) a sealed box; the whole point of an all-in-one system like an iMac is that it can be treated as a sealed box by the manufacturer, the dealer and the customer. Taking the same money from one group of customers as from another, and not delivering the same product as expected by those customers, should not be allowed in any country that has a functioning consumer-protection legal framework. Certainly not in the US, UK or EU.

    Apple need to immediately and publicly explain to the public (who include, after all, their paying customers) what the dimensions of this variation are: how wide the performance difference, how many customers are affected, and what plausible justification, if any, they have for not disclosing the variation to customers at or before the time of sale. Most importantly, they need to lay out and implement plans for making good their existing customers who have received less than the “full Mac experience,” as well as how they will ensure that such petty shenanigans will not recur in future.

    They should do this not just because it’s the right thing to do, but because by not getting out in front of this, they invite every consumer-protection and market-regulation authority in every country where Macs are sold to open up their own investigations, and eventually bind Apple to a patchwork of varying, even conflicting, remediation.

    Come on, Apple. You’re better than this… aren’t you?

Previous post:

Next post:

>