Network Computingwww.networkcomputing.com

That Was Then At that time, a good microprocessor ran at about 8 MHz. As a student, you usually could put together a project running your microprocessor at about 1 MHz or 2 MHz. To go much faster required designing and producing a printed circuit board--something beyond the means of a student project. Nonetheless, it was very instructive to build your own microprocessor-based system, burn your program into a ROM chip, and then try to figure out why it didn't execute right when everything was plugged together. Was it the software? Maybe it was the hardware. That was the fun of being an engineering student. Designing the thing was sort of enginee ring, but troubleshooting it, that was the real stuff.

Things are, needless to say, a little different today. There is more cache memory on most microprocessors now than we had for an entire class of projects in the '80s (64 KB was a huge amount of RAM for a student project).

This Is Now As fun as it is to chat about the good ol' days, the future--and speculation on where it's going--is where the real action is. The SIA (Semiconductor Industry Association) has just finished doing that. But before I relay some of its crystal ball findings, let's quickly look at where we are today.

Microprocessor speeds run commonly between 200 MHz and 300 MHz. For those MPUs destined for servers, the typical onboard cache memory is about 1 Mb. Memory chip density is as high as 256 Mb per chip. Slap eight or so of those on a DIMM card and you've got a cool 256 MB per slot. For purposes of comparison, my first hard drive was only 10 MB; on the other hand, my last PowerPoint presentation was that big. All of thes e chips run at 3 volts or so. That's important because the lower the voltage, the less overall power consumed. Low voltage not only saves batteries on mobile systems, it also saves cooling on other systems.

What's In Store Where does the SIA see the chip world going? Well, the association figures that it will be able to produce 256 Gb memory chips by the year 2012. My last hard drive was 4 GB; it makes me feel a little better about having bought that 10 MB drive way back when. It also makes me wonder what size my PowerPoint presentations will be a decade into the next millennium.

Microprocessors won't be lagging. They will incorporate as many as 1.4 billion transistors (the 8088 used in the original PC had around 100,000). Transistors in 2012 will run at 10 GHz, according to the SIA, but the overall MPU will run only at 3 GHz. Just a thought here: Your microwave runs at about 2.4 GHz (the wave length needs to be about the size of a water molecule, so that the oven can get those molecules hoppin ' around), maybe Intel will start making solid-state microwave ovens for us sometime next century. They'd probably be just large enough to hold a cup of coffee. If you make instant coffee in it, you'd conceivably move backwards through time. It probably really won't work, however. Chips are expected to use voltages as low as 0.7 volts by then. That's probably not even enough to warm a cup of coffee.

The moral of the story is that there will be no slowdown in the rate of computer speed and memory density increases for at least the next 15 years. It also means that the applications for computers should advance at least as much as they have since the days when we first used DOS 2.1 and Lotus 1-2-3 as our most powerful applications. The computer will be fully capable of doing greater things. The ones to feel sorry for (if any) are the software guys. They have a near impossible task ahead of them. But that's a story for another issue.

Art Wittmann can be reached at awittmann@nwc.com.