OK, you’ve read the hoopla and the hype.
Here’s the other side of the story.
Do you know what the core problem with socket 754 processors is? They’re great processors, for OEM machines. If you’re not going to overclock and stretch the envelope, and you want a bit better bang for a little less buck than a PIV box, the Athlon 64 fits the bills pretty nicely. An Athlon 3000+ would make a nice Dell box.
But I don’t think that’s exactly what the average reader is looking for.
This is what I’m afraid is going to happen.
For this audience, all this talk about Athlon 64 3400+ isn’t really about getting you to buy a 3400+. It’s to get you to buy any A64, soon.
The Athlon 64 3000+ is priced at a bit over $200. People are chomping at the bit to do something new. So they buy the “bargain” solution (and think themselves very smart for doing so).
Three months from now, we’ll have the Athlon 64 3000+ morphed into a socket 939 chip. It will probably get called a 3200+, and it will likely be priced at about $280, or about $60-$70 more than the 3000+.
Since socket 939 mobos will be better optimized for desktop apps, it will do roughly 5% better than the 3000+, and when you take into account the extra cost of a dual-channel mobo and two sticks of RAM, it will cost about $100-125 more than the single-channel platform.
Many will look at this, and again, go to the “bargain solution” and AMD will happily get rid of its broken “big” Hammers/soon-to-be relegated-to-the-bargain-basement socket 754 platform chips at a high price.
Towards the end of the year, we’ll see 90nm processors come out, and they’ll all be socket 939. Socket 754 owners will notice at that time that not only will there not be a 90nm socket 754 replacement, but that the new socket 754 replacement is a downgrade on what they already have.
So the socket 754 people will then lay out again for a new CPU, a new mobo, and two channels of memory (if necessary), to find out that AMD plans to go to DDR-II about six months after their belated purchase.
We know all this today. AMD isn’t even trying to keep this a secret, they’ve said so in their roadmaps. What is really happening here is that AMD knows and has acknowledged it made a mistake thinking they could get away with a single-channel memory solution for Hammer for mainstream/performance computing. All this hyping is an attempt to get you to pay for their mistake.
We don’t think that’s a good idea. To us, it seems like AMD wants to play impatient overclockers like a fiddle: Buy now, buy later.
I have to give them big marks for chutzpah, and this is going to be hard to beat for the Slick Willy Award in 2004, but that’s not exactly a coveted award. 🙂
I’ve been around this place for years. I know how many, maybe most people actually buy these things. They think a bargain is a low price today. They don’t want to think a lot about a purchase, and they especially don’t want to look ahead. Very often, they’ll buy something not because they need it, but because it will make them feel better about themselves, make them feel better than others.
So when a bunch of people tell them, “You know, you’d really be smart, have the best computer on your block and get one over on the Man if you bought this today,” they fall for it. How couldn’t they? They’re defenseless, and the Man is pressing all their right buttons while letting them think the opposite.
Is this deliberate? I don’t know, but even if it isn’t, your wallet ends up just as empty as if it were.
Let me put it another way. If I worked for AMD marketing, and needed to get rid of a bunch of surplus processors, this is exactly what I would do.
This is why we object to these CPUs. It’s not that we’re against the CPU or Hammers; we’re against the fiddle-playing, intentional or not.
Do others do it? Of course they do; it’s practically a principle of advertising these days. That doesn’t make it right or good, or mean you have to fall for it, anytime for anyone. It’s the practice we’re against, not the specific practitioner.
What we think is a better idea, even for the impatient ones, is to wait for for socket 939, make sure the socket 939 mobos will work with 90nm chips, buy that, and then upgrade the processor to 90nm when appropriate. You’ll have about a year of use before DDR-II is likely to show up.
If the initial generation of Hammers offered a huge performance difference over current Athlon XP systems, there would be a stronger case for buying right away. They don’t. The improvement you can expect on average is only about 20-25%. Someday, you’ll get an additional boost from x86-64, but they’ll probably be few and far between for 2004.
If you’re going to lay out $325 or $450 to get into Hammer, you might as well get the most you can for your outlay.
There’s also a third possibility . . . .
We suspect they’re going to be dogs, especially the socket 478 versions, but we don’t know that for sure, yet. However, a Prescott running at 4GHz ought to be able to edge any AMD Hammer in its price range (granted, without any likely x86-64 bonus down the road).
It says a lot about how screwed things at least appear to be at Intel that we can’t even presume the initial ones will hit 4GHz (which would normally be a “Duh” after a process shrink, even initially).
The socket T Prescotts (at least the later ones) ought to do better than the socket 478s, but the total cost of ownership on that will also include a PCIExpress video card and DDR-II, which will hardly make it a bargain upgrade.
And, like the first generation Hammers, the performance improvement over current mainstream products is if anything even less compelling than with a Hammer upgrade.
You can see we’re hardly dripping from enthusiasm over this option.
We don’t think Prescott is going to or ought to induce a lot of AMD fans to go blue, but just the thought that Intel will arguably offer the “bargain” performance upgrade shortly will make the blood of a lot of AMDers run cold.
We think it would be a very good idea for AMD to break their “no less than 2GHz pledge” and come out with a lower speed socket 939 Newcastle to at least match the price of the 2.8GHz Prescott. If Intel can afford to make one at that price, AMD certainly can, too.
The Prudent Thing To Do
If I were you, and had a relatively current AMD/Intel system, this is what I would do:
1) See how socket 478 Prescotts do overclocking-wise in February, and unless the initial ones do 4GHz easily (which we wouldn’t bet on), cross them off the shopping list.
2) See how socket T Prescotts do, probably at the end of March, and look at the overall cost of a socket T upgrade, especially for memory. Again, for most people, they’ll have to do much better than expected to earn serious buying consideration.
3) See what Intel plans to do about x86-64. In all likelihood, they’ll announce it for Tejas at the end of March. That would be a good reason to buy Tejas in 2005 rather than Prescott today.
4) Now that we’ve gotten Intel out of the way (in all likelihood, you’re not going to like the answers to questions 1-3), look at how the Newcastle socket 939s do (they should also come out around this time). A socket 939 2GHz Newcastle will probably perform roughly on par with a current Athlon 64 3200+, and probably a little better on games/apps that are memory bandwidth intensive.
5) Make absolutely sure that socket 939 supports 90nm CPUs, because that’s the main reason why you want to buy this.
6) Look at the cost of the system, and ask yourself if it is worth the price for roughly a 25% improvement.
7) If you’re not especially hurting, you may benefit by waiting until the fall to see whether or not AMD can make its “same price as Intel” stick. You may end up with a bargain. Even if you don’t, odds are buying just one 90nm processor will cost less than buying a 130 and a 90mm chip.
Eyes Wide Open, Or Shut?
We don’t feel that our job should be to encourage, entice, or excite you into buying things. There are plenty of others who do that.
We see our job as being to at least inform, at most to advise. We certainly base our advise on certain principles. We believe that there are better products to buy than others, and better times to buy them than others. This means it is better to look ahead, and plan your purchase rather than buy on a whim, or when somebody else has somehow gotten you hot and bothered enough to do so.
That’s pretty good advise for anything in life, and if you don’t think so, odds are you’ll eventually learn the hard, expensive way.
That doesn’t mean you need slavishly follow whatever we say. Not everybody is in the same situation, and sometimes the calls we make are pretty close.
In this case, some of the people reading this will say, “OK, so I’ll buy a socket 754 now, and another one nine months later. Big deal.” That’s fine by us. So long as you know what you’re probably getting yourself into and accept that, we’ve done our job.
What we really want people to do is think before they buy, and consider both the pros and cons. We emphasize the cons mainly because almost everyone else are hyping the pros.
What we don’t like to see are people swept up by a lot of manufactured emotion and excitement into wasting money they often don’t have a ton of. When you let your emotions make your decisions for you, you lose control to those who can stir those emotions and thus manipulate you into doing what they want while letting you think it’s your idea.
You’re spending money on a computer, not a girlfriend. Do you get excited about buying a screwdriver? No? Well, when you’re thinking about buying, keep telling yourself that. “It’s a screwdriver, not a girlfriend.” Maybe that will help.
If you find this approach terrible and horrible, ask yourself this question. Are you getting excited about getting a computer, or are you getting excited about being excited? Are you really buying a computer, or an emotional high?
Yes, this is very hard to do. No, you’ll probably never completely be able to control it. But in the long run, the more you think and the less you feel when you buy something, the better you’ll get at it and the better off you’ll be. You’ll be able to afford more good things you’ll be able to buy and truly enjoy, knowing that you really made the decision and that you’re in control of your wallet.