Brian's Brain

Yesterday, Intel unveiled its Centrino 2 (aka Montevina) CPU-plus-chipset platform, and as such I thought it'd be a good time to update you on where I stand on the notebook computer front. As a reminder, back in late May my long-erratic MacBook 'threw a bit' (or a few) in its Boot Camp-created NTFS partition while I was overseas, rendering me unable to boot Windows XP on the system. Fortunately, once I got home, I was able to resurrect my dependable Dell Inspiron 700m (complete with a brand new LCD and keyboard) in fairly short order. And I've been using the 700m ever since, though not without an occasional wince and gnashing of teeth, as I think back to the comparative responsiveness of its dual-core and otherwise more advanced (albeit currently crippled) predecessor.

The Inspiron 700m platform, after all, is four years old. My particular unit runs a 'Dothan'-generation single-core 1.6 GHz Pentium M processor, along with an 855-generation chipset (aka the Carmel platform), connected to 1.5 GBytes of DDR-333 SDRAM and a 60 GByte 5400 RPM PATA HDD. Basic computing functions (email, word processing, no-Flash web browsing) are tolerable, but anything multimedia-inclusive brings the system to its knees. And given my editorial beat, you can probably imagine the frustrations these performance limitation cause. Not to mention the mildly irritating flaky microphone input and lack of integrated Bluetooth.

In retrospect, thinking back on my multi-year Windows-native-on-Apple-hardware experiment, there was just something not quite right about the system-to-HDD interface. Hardware? Software? Both? I dunno. But in addition to the definitive infraction that occurred while I was in Taiwan, I'd previously noted plenty of warning bells; random corruption of my Outlook database, for example, or auto-launches into ScanDisk for NTFS repair on Windows boot. Applications also would occasionally refuse to run until I re-installed them, with Windows bafflingly reporting that they weren't 'valid EXEs'. Etc...

I'm hesitant to definitively point my finger at Apple's Boot Camp drivers (or, perhaps, its hybrid GPT/MBR approach, or its translation scheme for running a non-EFI-supportive O/S on an EFI foundation); perhaps a flaky HDD is the root cause of my sporadic woes. After all, even though I rarely ran OS X on the machine, I had at least one notable partition corruption-related crisis related to Apple's operating system, too.

Nonetheless, I'm done with Boot Camp, with the exception of using it as a reference point for some upcoming benchmarking work. To wit, I'm dabbling with running a virtualized copy of Windows XP under OS X on the MacBook; I'll ultimately test both Parallels and VMware, along with full hardware virtualization via Microsoft's Virtual PC for Mac (running x86-targeted Windows XP on a PowerPC G5-based Power Mac), and application-specific virtualization using CrossOver and its open-source Wine foundation. The results of all of this experimentation, data collection and analysis will appear in my October 2 EDN cover story.

Will virtualized XP be sufficiently speedy, reliable and feature set-robust for me to rely on the MacBook for everyday Windows work? I won't know the answer to this question for several months and, with past history as a guide to potential sub-par future results, I've decided not to wait to find out. Two days ago, I took the plunge on a system I've long been eyeing; Dell's highly rated XPS M1330. Ironically, this decision occurred coincident with Intel's press event in San Francisco (which I was unable to personally attend) and, ironically too, the XPS M1330 I bought is two platform generations behind what Intel publicly unveiled yesterday. Am I crazy? Maybe, or maybe I'm just more fiscally motivated than feature set tempted. Read on for details on my discernment process, along with a Centrino 2 analysis.

Continue reading with Part 2, 'CPUs and Chipsets: The Fiscal Upside To Going Downscale'...