Some Precautions On Conroe . . .

No surprise in this corner, but there’s a million websites showing a million stats indicating that Conroe wallops the AMD competition.

We’re not going to dwell on that, but just make a few points and observation that otherwise might be overlooked:

1) Don’t be penny-wise, pound-foolish If you understand nothing else about Conroe, understand this: To get maximum benefit, you’re going to have to be loose with the wallet.

XBitLabs has an excellent article on overclocking the low-end $183 processor. It emphasizes that a lot of performance got left on the table because of the low multiplier and the FSB limitations of the motherboard.

While matters may improve a bit with better motherboards, it’s unrealistic to expect mobos to give you 550MHz+ FSBs anytime in the foreseeable future. Given that, we think that it’s penny-wise, pound-foolish not to spend about $30 more for the E6400, and for maximum potential, bite the bullet and pay another $100 for the E6600.

True, Conroe prices will probably be inflated to begin with, but there’s another excellent reason to wait until they settle down.

2) Video problems Conroes have a video card problem: they need a PCI Express video card. If you have one, then there’s no problem and you can skip this part.

If you don’t, and you like gaming, then you have a problem. If you don’t have a PCI Express video card, odds are you haven’t upgraded for a while, and odds are you’ll be in no mood to upgrade whatever you buy anytime soon.

If you expect your video card to last two-three years, odds are you’ll end up upgrading to Vista sometime in that period and start playing games that will at least be able to use DX10 to enhance the game.

DX10 video cards don’t exist yet, and won’t until the fall, likely late fall. Additionally, there’s some indications that overclocked Conroes simply outrun current video cards.

If you want to buy a video card now, and don’t mind buying another one 12-18 months now when Vista and DX10 become appealing to you, again, no problem and skip the rest.

If you do mind doing that, though, you might want to wait a few months until such cards come out. By then, Conroe prices will probably have settled down as production increases, and you can buy knowing that some hot game won’t obsolete your video card.

3) Don’t panic! Really, truly, you don’t have to be the first kid on your block to get one of these things. Right now, I suspect there’s a lot of people just salivating at the thought of making extra money from the suckers self-controlless enthusiasts who think paying extra makes them better people.

Really, hopefully sometime late in the century (or maybe even next century), when you’re on your deathbed, and people ask you, “What did you regret not doing in your life?” the older, wiser you is not going to say, “I didn’t buy a Conroe back in August 2006.”

Take the money you would have otherwise spent paying above retail price and put towards a better video card or more/better RAM. You’ll be happier in the long run, and more importantly, learning a little patience and self-denial until the time is right now will make you a better person later, at least a better buyer for bigger, more important purchases later.

4) Make AMD show you the money If you have a socket 939 system and are thinking, “Well, I’ll just buy an X2 when they become dirt cheap,” skip the rest of this section. I’m not talking about you; that’s a reasonable step.

However, if you bleed green, and want to spend more than a trifle equipping an AM2 system to show faith, don’t. At least don’t until AMD shows some real sign your love is not unrequited.

How can they do that? Well, what did Intel do with Conroe? They demonstrated what they had months ago (and seem to have leaked more than a few CPUs after that to keep the fires burning. Even some Intel quad-cores have shown up for pioneer testing).

If AMD wants you to wait for them to catch up, then they can do exactly the same thing Intel did, and you shouldn’t give them your money until they do. They ought to love you at least as much as Intel did, and if they won’t, maybe this isn’t love.

Right now, it looks like AMD got caught with their pants down, won’t be able to pull them up for sometime to come, and are coming up with last-second expedients like 4X4. A large part of this appear to be the usual suspect: getting SOI to run quickly. IBM is having huge problems making fast Cell processors for the PS3; IBM itself admits that they’re only getting 10-20% yields on Cells, and Sony may have to drop Cell’s speed from 3.5 to 2.8GHz.

True, Cell is a more complex processor, and SOI probably isn’t the only or maybe even major reason for those specific problems, but given the rather, uhh, low speeds AMD is promising from 65nm processors, it would be foolish to think SOI isn’t part of the problem.

You may want to think AMD has a secret weapon they’re just waiting to unleash at the right time, but in the eight years or so I’ve been watching this, the only “secret weapon CPU” I’ve ever seen has been Conroe itself, and what did Intel do? They showed it well ahead of time.

Sorry, but thinking AMD has secret weapons armed and ready is about as much wishful thinking as Germans during WWII thinking Hitler had them.

If AMD wants you to buy a system now on faith that they’ll eventually give you a world-beater, they ought to give you more than vague words that guarantee nothing specific to show that your faith isn’t blind.

When people give you just vague promises that promise nothing specific, you ought to think “they don’t have the goods, at least not yet” until they show you otherwise. It’s their job to show you, not your job to believe.

If they love you, then they’ll show you. Make them justify your love.


Be the first to comment

Leave a Reply