Computing Event of the Year
The End of the Era of Speed: The age of the era of unfettered speed increases ended this year, symbolized by Intel’s unannouncement of a 4GHz chip.
However, this is not an Intel-only problem. No one has a clear lane any longer. It’s not that chips won’t get faster, but they’ll get faster a lot more slowly than in the previous twenty-five years of the PC era, and the costs will rise.
This bottleneck is likely to push the PC industry into all different kinds of directions over the next few years.
Intel Pushes It Until It Breaks For the second time within five years, Intel pushed a design until it broke, and for the second time in five years, they’ve basically been caught with their pants down. This is getting to be a habit.
Intel was king during the Era of Speed. Can it retain that title as the definition of what is a good CPU morphs into something new and unknown? There’s good reason for doubt, especially when . . . .
IBM Starts Rumbling: While the media noticed IBM’s sale of its PC-assembly business, it did not generally notice IBM sticking its fingers into just about every computing pie. XBox2, Playstation3, collaboration with AMD, IBM positioned itself and its coalition of heavy-hitting willing allies to mount a challenge against Intel for King of the Hill the next half of the decade.
Overclocking Product of the Year
None. Hey, they didn’t give out the Nobel Peace Prize too often during World Wars I and II, so why give a prize to something that isn’t really there?
In 2002, we gave the prize to the Intel Northwood 1.6A processor. In 2003, we gave it to the AMD Thoroughbred B processor. Nothing in 2004 compared to these products, simply because to win this award, a ton of people have to buy the product, and there was no product like that in 2004.
The only real contender for the prize were the low-end 90nm Hammers, but they came a little too late, cost a little too much, and offered a bit too little performance increase to make the typical AMD fan jump on the bandwagon.
Other contenders like the latest generation of video cards came too late in the year (and often were half-vapor even after they “came”) to make a major impact. Others (like DVD burners) were at least somewhat crippled by subsidiary factors (the cost of dual-layer media).
It’s pretty safe to say, though, that all the abovementioned will ripen into legit contenders next year.
Dog of the Year
Prescott Underperforming, overheating, do you need any more? And all Intel’s horses and all Intel’s men, couldn’t put Prescott together again. Probably the worst CPU Intel has ever released.
Winner of the Year
Nobody: It’s obviously not Intel, AMD improved, but hardly set the world on fire, ATI finally reaped big gains in market share at the expense of nVidia, but nVidia got better and probably has the edge the next go-round. Nobody won big-time.
Loser of the Year
Us Delay, delay, cancellation. Delay, delay, cancellation. Outside of a (relatively) few number of people buying Hammers, there was nothing significant to do this year. Even towards the end of the year, the scene was fogged in with all the vapor(ware) in the air.
Intel The reasons are self-evident.
Slick Willy Award
AMD: AMD didn’t have a bad year. Opterons grabbed a small but growing section of the server market. They caught a break when the PC market revived late in 2003, and sold enough Athlon XPs at decent prices while topping up revenues enough with a handful of high-priced Hammer sales to basically break even early in 2004. Later on, they sold more Hammers at lower prices, again to basically break even for the year.
For most of the year, they had real problems making Hammers, but finally managed to straighten that out (we think) with some help from IBM. Hammer ought to really launch in 2005.
That’s a lot better than losing hundreds of millions of dollars and not being able to ramp up your latest product, but it’s hardly a roaring success story.
Yet AMD has managed to convince most of the world that it is, and that it has Intel on the ropes, or close to it.
That in-and-of-itself would be pretty impressive, but what makes AMD’s PR performance truly outstanding is managing the feat while essentially giving the finger to anyone and everyone trying to peer behind Oz’s screen. It’s like Donald Rumsfeld winning a Bill Clinton Charisma award from the press.
It was almost a Steven Jobs level performance.
What To Watch In 2005
Getting Hammered: For better or for worse, whatever happens in the CPU field is going to happen with Hammer. Intel is essentially knocked out of the box for this audience in 2005; all it can do is make moves that AMD will have to respond to.
The one Intel move most likely to affect AMD buyers is adding x86-64 capability to their entire desktop line, most particularly Celeron. AMD will have to then make socket 754 Semprons x86-64 capable, too. If such Semprons (presumably second-generation 90nm) can crank up as high as their big brothers, and the performance penalty isn’t too great, these are likely to become quite appealing chips to some AMD fans.
What we don’t know is how many AMD fans are likely to buy this beast if/when it arrives. This could be a minor overclocking front, like Duron overclocking was a few years back, or it could become the Celeron 300A of 2005.
A lot of that will depend on the pricing and configuration of low-end 512K “real” Athlon 64s. If socket 754 64s go next-generation 90nm, too, and end up selling six months from now for less than $100, that could well be the overclocking weapon of choice instead, but if that doesn’t happen, we think Semprons will look a lot better to people than they do right now.
We think we’ll see two groups of Hammer overclockers, a high-end of socket 939ers, and a low-end of socket 754ers. How many will be in either camp is uncertain, but I guarantee they’ll fight a lot about it. 🙂
The overall market? We expect AMD to increase its marketshare from the current 16-18% into the low twenties, not good for Intel, but not terribly bad, either. We think AMD will choose higher prices over maximum marketshare, and unlike the Thoroughbred era, we think they’ll be able to avoid the bargain basement this time around. By the second half of the year, we expect AMD to be making a couple hundred million a year making CPUs.
The Screens Come Tumbling Down: Expect prices of larger LCDs and other advanced monitor technologies to tumble in price this year, from a lot too expensive to just somewhat too expensive. LCDs should also start providing decent response times for those who want to make things move a lot on their screens without thinking their monitor is haunted by ghosts.
While this development will have more impact in the TV area; the difference between a monitor and TV will continue to blur as PCs aim more towards the home entertainment field, and people begin to warm to the idea of big-screen watching and computing.
The DVD Burner Arrives: Enthusiasts will yell this day is long past, but DVD recording is still not a brain-dead activity. That will only happen when dual-layer media drops under a dollar a pop, and DVD copying becomes just that rather than one of remastering, compiling and burning.
At that point, DVD burning will become mainstream, which horrifies Hollywood even more than W’s reelection, which leads to . . . .
Make It Go Away, Just Make It Go Away
2004 wasn’t just a bad year for technology. It was an even worse year for politicians trying to figure out what to do about it.
They don’t understand the phenomenon. Most don’t want to understand this, and the few that want to haven’t or maybe even can’t understand this.
I suspect the average legislator views P2P and digital copying much like he or she views North Korea/Iran and its nuclear weapons. It’s easy to be against the idea, and you don’t want to do nothing about it, but when you look at the “somethings,” they might lead to some seriously bad things happening, because you don’t think your opponents are screwed on all too tight. So you drift along and hope the problem goes away or gets somehow taken care of with no real effort on your part.
The one big difference between the two is that the opponent is not somebody named Kim, but your (and your constituents’) kids. They’re not exactly rational on the subject, either, and they’re doing harm, but you don’t want to ruin their lives over this, either. That really makes you want somebody else to take it off your hands.
The only possible solution out there that offers any chance of solving the problem to the legislator’s satisfaction is hardware-enabled DRM backed by some form of legislative mandate. By the end of the year, the technology companies should be ready to tell legislators, “Don’t worry, we’ll take care of this.”
Maybe. It’s only a partial solution even if it works perfectly, and the content providers aren’t going to like that at all. The kids and their Pied Pipers certainly won’t like this, and even the people providing the solution would much rather not.
It’s unclear whether legislators understand the issue enough to realize that it’s really a choice between DRM and/or establishing the Copyright Police. There are no other alternatives. It’s extremely doubtful that they have a clue that they’ll also need to revise copyright law for the digital era no matter what they choose. They certainly aren’t getting any reasonable guidance from those most interested in the subject.
The inability of governmental systems to cope with or even comprehend an increasingly important cyberspace does not speak well for the future.