Computer Speedup Is Just a Chip Away
THE morning was beautiful. Sunlight reflected off the thick snow. A fine day, I thought, to upgrade my computer by exchanging my 386-class microprocessor for a faster, 486 ``upgrade'' chip from Cyrix Corporation.
The manual was reassuringly thin. If Cyrix could explain how to replace a microprocessor in 28 pages, it couldn't be that complicated, right?
I breezed through the first five pages of the manual. Then I ran into Page 6 - and the productivity paradox.
The first challenge was clearing a path to my microprocessor. I pulled out my hard drive and the CD-ROM player on top of it. That gave me enough room to slip a chip-pulling tool under one side of the microprocessor. But several electronic gizmos blocked access to all but one other corner of the chip. I squeezed the tools together, the way the manual said. Nothing budged. I squeezed harder. Nothing.
I glanced down at the manual's diagram and realized my mistake. I was trying to dislodge the chip socket instead of the chip sitting on top of it. I readjusted the tools and pulled off the 386 microprocessor.
The Cyrix upgrade chip is a clever bit of engineering. It fits into the 386 socket but its instruction set mimics speedier 486 microprocessors. It has an on-board cache that allows quicker access to crucial information. The chip is also clock-doubled, which means it runs twice as fast internally as the rest of the system. The result: My 386 20-MHz system should start acting more like a 486 40-MHz system. All with just a $349 upgrade chip.
Most of its buyers are end-users like me, rather than corporations, says Paul Pascarelli, Cyrix upgrade marketing manager. One source says the company expects to sell ``a couple of tens of thousands.''
My first mistake in installing the chip was handling it too gingerly. When the hard drive accidentally bumped it during reinstallation, the chip popped out. I pressed it down firmly the second time.
Then I couldn't boot (start) the computer. The screen flashed a message that my hard drive had failed. I checked all the connections. I called Cyrix's toll-free number, and a technician suggested I recheck the connections. (While on hold, the company's recording said the upgrade chip could be installed in about 15 minutes. I was going on two hours.)
After lunch, I found the problem. My hard drive needed TWO cables, not one, hooked to it. Reconnected, everything worked perfectly. Well, almost perfectly. I also had replaced the CD-ROM drive with a tape-backup drive, which the software couldn't recognize. That stumped me for an hour and a half until I cleaned out the dust from inside the tape-drive. Bingo! At 3:30 p.m., I was up and running again.
This is the productivity paradox. The new chip allows my programs to run, conservatively, 15 percent faster. (Cyrix claims a 100 percent increase or better.) If I spent half an hour a day using the machine in a way where that speed counts, the chip would save me 4-1/2 minutes daily. But it took 3-1/2 hours to install it and the tape drive, meaning it would take 47 working days just to catch up. Was it worth it?
Technologists argue this endlessly. For all the money America's service sector has poured into computers - an estimated $860 billion in the last 10 years - their productivity has scarcely improved. One group says computers don't make people more productive. A recent National Research Council report argues that the productivity measures are missing the gains. A third camp, to which I belong, says it takes time for people to figure out how to use this stuff.
It has taken me weeks, even years sometimes, to find more productive ways to use software. I have made time-consuming mistakes installing new hardware. But I'm further up the learning curve and faster too.
If my figures are right, I'll be ahead on productivity when the winter snows begin to melt.
* Send your comments on this column to CompuServe (70541,3654) or Prodigy (BXGN44A).