Drought or Deluge

Add Your Comments

We have two groups of people out there, those who need machines that have to work, and those who don’t.

If you’re playing a game, you can live with a machine that fails on you every once in a while. Sure, that will get you mad at the time, but normally, it’s no big deal. There’s always the next game.

If every crash costs you money, though, you can’t afford to have a machine that does that once a day or so. You can’t afford to have CPUs blow up on you. You can’t afford to hope that you’ll get a “good” motherboard rather than a bad one.

That’s the biggest problem we face right now with the Athlon systems. They just aren’t reliable enough for the second type of person.

I don’t read about Intel CPUs blowing up left and right. That was and is a great rarity. It happens too often with the Athlons. Granted, most of them are caused by users, but I’ve seen too many cases of these things dying for no good reason not to believe the reliability standard is a bit lower than what we were used to a few years back.

That’s hardly the major problem, though.

Via is turning into a four-letter word for me. I feel like a nursing home attendant at times with these machines, continually feeding my machine medicine to fix one thing or another and keep my patient alive.

Put in this driver; stick in this fix, put in that fix sometimes, put in another fix another time, and if you don’t perform the ritual exactly right, sometimes you’ll have to start all over again. I’m installing a motherboard, not a bishop.

Nor is this a matter of hanging out on the edge. Most of that medicine is meant to get the machine to work right at any speed.

Even if you have or even know to have all the necessary drugs, that still doesn’t guarantee you a healthy patient. Too many of these boards are just bad. Not most of them, too many of them.

To me, the issue is not “Does Via make a quality product” or “Will Via make a quality product?” it’s “Can Via make a quality product, and does it even want to?” I think the answer is “No.”

Remember, this is a company that had 5% of the chipset business until Intel decided to give it away. You don’t expand tenfold in two years without big problems. Add to that AMD changing standards every five minutes, and you get what we got now.

None of this may bother the hotrodders at all, and that’s perfectly OK. If it doesn’t bother you, I’m not suggesting it should start.

But if AMD ever expects to get out of the hobbyist/consumer ghetto and wants to break into the business world, they have to do better than this.

Look at this from the perspective of an IT person. When it comes to desktop systems, he has one and only one top priority: no headaches, or at least no new ones. He couldn’t care less if your Intel system is 10 or 15% slower than an AMD system; he’s not using it. He couldn’t care less if your Intel system costs more; he’s not paying for it.

What he cares about very much, though, is how much effort he’s going to have to make and how many headaches your computer system is going to cause him.

Right now, the AMD solution means more effort and more headaches. That’s not a winning combo.

Right now, there is a tradeoff buying AMD platforms. Higher speed and lower cost in return for lower reliability, and that’s a perfectly acceptable tradeoff for many, but not for all.

Right now, I can’t say to someone whose hardware must work all the time, “Buy AMD.” I just can’t.

Now most people aren’t really in that situation; reliability is desirable, not critical, and other factors are more important. That’s fine. You may have an AMD machine that does everything you want flawlessly. Good for you, but you are not the universe.

Like many other people, I just want a machine that works without constant care and feeding and worrying, and I’m not getting that so far.

We’re not going to give up looking at AMD by any means, but we’re going to start taking a closer look at Intel systems over the next few months to see whether or not their new offerings are dropping the ball on reliability or not, too.

Email Ed


Leave a Reply

Your email address will not be published. Required fields are marked *