Yes, I know that headline is provocative, and I admit it is just an estimate, and your mileage may vary. But read on to fully understand...
There has been some interesting stuff around the web over the last few weeks about whether the new Core Duo chips have a bug which causes excess battery drain any time a USB device is connected (mouse, printer, card reader, camera, iPod, etc). This would theoretically affect Macs as well as Windows machines.
Now, however, the problem has been clarified in some thorough tests by Anandtech. The problem is a real problem, but turns out to exist on other Intel platforms (eg Pentium M) and may also exist on AMD systems too. The good news for Mac users (and Apple) is that the bug is not in the chip or the chipset, but in Windows XP. And there is a patch for it from Microsoft, though not a very nice patch (edit the registry), and one that doesn't work if the computer is put to sleep and woken up. The problem appears to be due to the MS USB driver which is loaded whenever a device is connected and which stops the processor from dropping down into a lower power state.
The effect of the problem is to reduce battery life by between 15 and 28% on 5 systems that were tested. Devices with permanently-connected USB devices (eg a built-in webcam as found on some Acer models), will have this problem all the time. Even disabling the device driver does not work in the Anandtech tests.
There is no word on a permanent fix. But for Apple it does mean that design-wise they start out with a 20% or so advantage in power use in many real world situations. Of course other factors can come into play - other parts of a computer may use more or less power, and the OS can behave differently with respect to power handling. Battery specification also obviously matters too. It is also to be noted that no PC manufacturer will quote their battery life WITH a USB device connected, so specifications may not reflect those advantages. But assuming Apple hasn't screwed up in some other power use area, for those people wanting a laptop with a built-in camera or who expect to frequently use ANY USB 2.0 device, a Mac will likely offer a longer REAL runtime than comparable Wintel machines.
Update: Clarification of this bug over at InfoWorld with Microsoft comments. I would classify this bug as more serious than Microsoft does, and I would expect quite a few people and companies to be pretty upset with their approach (or lack of it) to dealing with it. While I expect many vendors to ignore the problem in their specifications (and many reviews will just quote from specs rather than test), a real-world laptop user will find a significant shortfall over what they expected.
No comments:
Post a Comment