Got some feedback on yesterday’s article about 64-bit, and as expected, it was generally negative.
The comments generally could be summarized as “well, it will benefit “us”, so why not?” and this illustrates the problem.
The computer industry isn’t “you” anymore.
Changing 100% To Benefit 5%
Going to 64-bit will certainly make a few niche groups happy. What Intel is basically saying is “We’re not going to change things for everybody just to make these niche groups happy.”
Does Intel have ulterior motives behind this? Of course. In the long run, they want to get rid of x86. In the medium-term, they want to provide an industrial-strength 64-bit solution (and make some good money from that) as a prototype for the future desktop chip. In many ways, Itanium is the Pentium Pro of this decade.
They have a much different vision of 64-bit than AMD has, which is essentially a quick fix to x86.
I will not argue the merits and demerits of the two approaches, that wasn’t the point of yesterday’s article or today’s.
The important point is that there’s a stark choice to be made. Intel is saying, “Here is where we stand. Prove us wrong.”
If you want cheap 64-bit for less than Itanium, your choice is clear. If Intel doesn’t want your kind of business, give it to AMD. Buy a Hammer, buy a toolshed of Hammers.
Understand something. Wanting x86-64 doesn’t make them bosom buddies with AMD, in fact, from the emails I’ve gotten, it looks like the opposite. These folks aren’t whining about not getting x86-64; they’re whining about not getting x86-64 from Intel. They don’t seem to want an AMD solution.
I suspect many are at least closet Intelverts. When the time comes, I’ll look at the product, and if 64-bit suits me better, I’ll buy AMD; I have no problem with that.
So why do they?
If not simply from anti-AMD bias, then maybe this is the reason . . .
Death Match: Joe Vs. You
In the one corner of the wrestling ring, we have a few 64-bit nerds, with a manager wearing the AMD logo.
In the other corner, we have a mob of Joes. Some are carrying six-packs, others are wearing suits, and the manager is wearing Intel blue.
The nerds are upset because the Blue Man is normally in their corner. They are really upset because in all their previous matches with the Joes, the Blue Man fixed the fights so that mob of Joes just rolled over and played dead as soon as the nerds touched them. Of course, the nerds think this makes them champs and say “What fix?”
But what they are really, really upset about is that they’re yelling, “Is the fix in?” to the blue guy, and he’s flipping them the bird. They look back at the green guy, and his facial expression says “Don’t look at me, you’re the champion wrestler.”
This leaves at least the few perceptive nerds with the sinking feeling in their guts that this match is real, and they’re going to have to pin all those Joes down without any outside help. Intel isn’t going to do the dirty work for them.
That’s the situation in a nutshell. Can you make all these Joes go to 64-bit without the blue guy fixing the fight? Can you pin at least enough of them down to make the blue guy think twice and toss in the towel?
Talk Is Cheap
Intel has made up its mind. All the slick talk in the world isn’t going to change their mind. Only sales figures will.
And not necessarily Hammer sales, either. Sales of Windows x86-64 and Linux 64-bit distributions. Sales of 64-bit applications and games.
In short, action, not talk, and not just action on your part, action on the part of the sea of Joes out there.
This has nothing to do with the merits or demerits of Athlon 64s or Opterons or x86-64 vs. Itaniums. It’s not even the usual AMD/Intel divide. Your average AMDer will or won’t buy a Hammer based on its price/performance, not because they love 64-bits. To them, that’s a fringe benefit. He won’t lay out a lot of extra money just to run 64-bit, that defeats his purpose for buying AMD.
In short, the average AMDer will run 64-bit only if it doesn’t cost him appreciably more to do so. If it does, he’ll either stick with 32-bit or say, “Warez Windows x86-64?” in certain places. 🙂
No, this is a test of power. Who controls the personal computer agenda now: the geeks or the non-geeks?
If the geeks control it, AMD can make a lot of hay for a couple years. If they don’t, so long as AMD doesn’t get delusions of grandeur and keeps giving people the usual financial reasons to keep buying their desktop processors, it will still probably be a net positive.
So Hammer can be successful enough even if x86-64 turns out to be a relative failure.
Intel is probably not doing what it’s doing as a test, but it’s an inevitable byproduct of their stance. Just how much power do the geeksters have on their own nowadays? And if the answer ends up being “Very little,” that will have big ramifications in the future in Santa Clara.
Today, there are two very separate types of computer buyers. There is a small minority who can legitimately say that nothing is ever enough, and the majority who can and are saying, “Enough.”
Up to now, the first has managed to get a mostly-free ride off the second. The cost of developing and making the technology got spread over a much bigger population, and at least the second group did get some noticeable improvement for their money.
But those days are ending if not ended, at least for a while, and group number two isn’t swallowing “more, more, more” so easily any more.
While Intel has its own reasons for doing this, if their customer feedback indicated that they needed to move sooner rather than later, they wouldn’t be talking about 2008 or 2009. At the least, we’d see Desktop Itanium Jr. well before then.
Intel is saying, “You want cutting-edge stuff, go pay for it and buy an Itanium” and I think some of these 64-bitters are going to be shocked when they see Opteron pricing. It will be a lot cheaper than Itanium, but not a lot cheaper than Xeons.
In the years to come, we are very likely to two markets emerging; one for the high-enders, and one for everyone else, with high-end prices that stay high-end for the high-end stuff.