Mac Pro indecision

I'm still trying to buy a Mac Pro. Actually, that's not quite true—I'm still trying to decide to buy a Mac Pro. Last month, I asked whether there were new models likely to be released soon. Steve's comment was spot on:
Yes, new models are on the way. The dual core Xeons are being replaced with quad core, so the new models will have 8 cores instead of 4.

Not long after that, Apple did, indeed, release a quad-core Mac Pro. However, and here is the first source of my current problem, they didn't so much replace the existing models with quad-core variants, they tacked a quad-core model on at the high-end. And, as far as I can tell, the low-end models didn't move in price—the entry-level configuration still starts at $A 3,999. Now, I don't know what I should have been expecting, but I know what I would have liked: a new machine at the high-end (of a three model spectrum), with the models below it shuffling down a price-point, and perhaps the lowest-end model just disappearing. If that had happened, I suspect I'd be typing this on a Mac Pro right now.

Steve also wrote:
So I would wait before buying one. Have another look at the situation when 10.5 comes out.
That seemed like good advice at the time, because at the time 10.5 was coming out mid-2007. Now, Apple has pushed Leopard back to October [I can't find a likely permanent link to that announcement at the moment]. I suspect that's still good advice, but it doesn't help my ever-increasing desire to upgrade my ageing Pentium 4 running a very creaky Windows XP install.


  1. Get a Mac Mini to keep you going for the time being. You may be surprised and find that it's all the computer you want.

    If it's not then it'll do for a few months, then shove it into your stereo cabinet when you're ready to update to a Mac Pro.

    With Front Row, you'll be able to stream videos from your Mac Pro to the Mac Mini, which then goes to your TV.

    (make sure it's the Core Duo Mini)

  2. This might factor into your value calculation:

    The memory bus does have a bottleneck, but it kicks in earlier than I expected.

    However, the testing methodology is suspicious as they're using single programs at a time. That would hit the same areas in the processing path, compared to a more typical mixed load.

    Even with that suspicious testing approach, it's 50% more performance for 20% more price, so it's doing a good job on the value curve.


Post a Comment

Popular Posts