What I Want For Christmas

First thing I have to do is move Christmas. Santa can’t deliver what I want by December, besides, you always get better deals off-season. 🙂

So let’s move Christmas to sometime next summer. That’s enough rest for Santa.

What do I want?

1) The fastest .13 micron CPU setup available: If that sounds vague, that’s because I’m not too sure what that will be.

Per processor speed, I figure the raw Mhz will come in somewhere inbetween 2.5-3.5Ghz.

On the AMD side, at least Thoroughbred had better be available by then, and Clawhammer might persuade me to tell Santa to lay off for a few extra months, don’t know yet.

On the Intel side, Willy2.

However, MHz does not a computer make, not anymore. The irony is that (outside Clawhammer, which is the wild card) the CPU won’t be the decisive factor, but the mobo chipset.

On the AMD side, we will likely have either nForce2 or something from somebody else which looks a lot like it. A big question is whether Intel will have an equivalent to nForce a year from now.

AMD’s HyperTransport (or equivalent) should be a given for AMD mobos a year from now. While Intel has its own competing standard, it doesn’t look like it will be around a year from now.

An obscure but highly relevant question is whether or not all the bandwidth delivered by the CPU can be handled by it. There’s no practical point having 5.4Gb/sec bandwidth if the processor can only handle half of that.

I suspect that by this point next year, comparing AMD to Intel is going to be an apples and oranges comparison. Each will handle certain operations a good deal better than the other; which is best for you will depend on what you do.

Dual-processors? It’ll be considered, though with my work habits, I doubt it will do me much good.

2) Very big, very fast, very cheap RAM modules: 1Gb (or even better 2Gb) PC2700 modules would suit me fine. The next jump in DDR speed won’t come until late 2003. Rambus has quicker modules in the works, but they don’t look they’ll get a ton better in a year’s time.

Hard drives are too slow now, and are going to get relatively worse. They are the bottleneck of systems today. I just can’t see hard drives being used in any serious system for other than backup or archival purposes ten years from now.

I grant you, this is not exactly an imminent crisis, but it’s inherently ridiculous to have Giga-Wiga-Hertz processors spending most of their time waiting around for rust.

At some point in the next few months, I’m going to try to get together at least a 2Gb rig and test to see just how much improvement I’ll get essentially shoving everything into RAM, for instance, running something like SiSoft with all of SiSoft in RAM. (I suspect Willies will like this a good deal more than Horsies.)

I suspect I’ll be somewhat disappointed in at least some of the results, but I already know who the major culprit is likely to be: the OS.

OSs/applications are simply not designed to operate with everything in RAM. MS acts like it owns stock in the hard drive companies.

Redmond gets even worse with its applications. If I wanted to run Word completely in memory, just how am I supposed to easily figure out what files I need? Required files are scattered all over the place: in the app folders, buried in the OS. I don’t want to have to play Indiana Jones or work for MS for five years to do this right.

True, there are programs that will let you create big RAMdisks, but that’s hardly an elegant solution, and it’s inherently wasteful; you end up with duplicated files occupying precious memory.

I want an OS that natively and intelligently lets you include as much or as little into memory as you want.

I grant you that this is premature for the vast majority of people, and hardly cost-effective at the moment, but even now, there are serious power users who would be delighted to have this now.

I think RAM-only functional systems (you’ll still have hard drives, you just won’t use them much) will be the wave of the (somewhat distant) future. Like LCDs were three years ago (though this one will probably take longer to become affordably practical).

The question I’d like to resolve, at least in my own mind, is to see whether it’s better to keep increasing the engine size, or widen the fuel lines.

Video: I like my dual 21″ Trinitrons just fine, and Matrox gets the job done for me.

However, running 3DMark 2001 with it is a sickening experience. I may complain about the extra cost, but one benefit of this nForce2 chipset would be to have a creditable gaming card built into the machine when and if I need it.

Hard drives I’m obviously not too big on hard drives, but I would hope to see 10,000 RPM IDE drives available by then.

Monitors: I personally don’t see myself with an LCD for a few more years, but I know that will be the replacement for the Trinitrons, and expect that reasonably big ones should be somewhat reasonably priced by then. I figure LCDs will start becoming standard starting about two years from now.

Other items: I have a 16X CD-RW, can’t see myself replacing that anytime soon. We’ve reached the point of diminishing returns. What am I going to do, save a minute tops for the rare occasions I burn one?

DVD burners will still be too expensive, and I can’t think of a legal reason why I might need one for years to come.

For networking, we probably will see some serious advances in two areas. I expect wireless routing to become pretty affordable within a year. For those who don’t mind wires, I think Gigabit Ethernet will become an expensive though not ridiculous alternative to 10/100. All these mobo advances will help that along, too.

Email Ed

Be the first to comment

Leave a Reply