Intel Quietly Stumbles Around

The last few months, critical attention has been focused on the AMD Follies than anything Intel has been doing. Not that they’ve been doing anything tremendous or stunning, just that they’ve been executing according to plan.

At least for the obvious things.

Then there’s the not-so-obvious.

Where Are the C1s?

On a reseller level, C1 stepping PIVs have not shown up below the 2.4GHz level, and even that is pretty iffy. Intel was still packing out B0 C1s as of a month ago.

Below that, the only way you can get a C1 stepping chip is to buy a computer from a big OEM, most particularly Dell. A number of people have actually done just that to get a C1 stepping 1.8A.

So the chips exist, and they’re being made; we’re just not seeing them.

I’m not all that surprised by the lack of 1.8As; Dell has been selling uber-cheap machines equipped with them, and I bet they’re selling like hotcakes. The 1.6A and 1.8A were meant to be primarily OEM chips, anyway, and only excess supply pushed them to the retail market.

What I am surprised about is not seeing these at the 2 or 2.26GHz level yet. Makes you wonder if the channels aren’t a bit clogged with older chips, a la AMD, or if Intel is in no rush to convert plants over.

Erratic C1 Performance

The database and forum reports on the C1s are a bit mixed. While these are usually getting to 3GHz, the early version of this seem to require a bit more voltage than expected to get there, and some just seem to fall flat on the face.

Interestingly enough, the few Dell 1.8As we’ve seen seem to do rather better getting to 3GHz with little additional voltage than their bigger (sometimes much bigger, as in expensive) brothers. It may well be that they’re more recent with a few more manufacturing tweaks in them.

If any of you have one of the Dell 1.8As, and can read the codes on them, could you send me a note telling me what they are? Thanks!

Where Are the C1+s?

For lack of a better name, that’s what I’m calling the variable voltage successors to the C1s. The rollout on these was supposed to be faster processors first, slower processors later. The projected rollout for the faster ones has been delayed almost two months.

Granite Bay Not Good For Gaming?

With a high price tag (in all fairness, it wasn’t meant to be a mainstream motherboard), Granite Bay was always going to be a hard sell.

It turns out that the Granite Bay chipset has a few video-related errata (see page 9).

It’s not clear if the first problem will show up in retail boards. The fix to this problem is supposed to be a board workaround, not sure if the earliest versions will have this.

The second problem, though, involves the AGP Prefetch Cache. It doesn’t work quite right, not right enough for Intel to tell developers to disable the feature.

That’s hardly going to help performance (though the performance hit is probably minor).


Granite Bay wasn’t meant to be a mainstream mobo. Springdale is. That means you’ll likely see lower prices, and some more features for your money: native serial ATA, higher official FSB support, a separate channel for Ethernet devices and maybe built-in 802.11b.

There’s some confusion out there on just what is what. There will be three kinds of Springdale boards; the integrated video “G” model; the low-end “P” model and the high-end “PE” model. The “PE” model will be the one with the 800MHz system bus, and will the mainstream version of Canterwood, which will be the workstation model.

Maybe Granite Bay Isn’t Such A Good Idea After All

None of these factors in-and-of-itself is a killer, but when you add extra cost, potential video problems, a new generation mobo coming not much later, and no cheap C1s or better around, maybe this isn’t the best choice in the world.

Getting Cheap

A couple years ago, the price point at which people bit at CPUs was about $250, and people really started piling on once the price crossed the $200 point. No more.

Whether it’s Intel or AMD, the 2400MHz level processor is not getting much attention in overclocking circles. It’s not that there isn’t interest in TBredBs or C1 PIVs; there’s plenty of that. It’s just that interest doesn’t turn into purchase at around $200.

What’s getting the most interest right now (though hardly a flood) are the 1700+ and 1800+ TBredAs. They may not get close to 2400+ levels, but the price is right.

I don’t think we’re going to see cheap and good from anybody for quite some time. Maybe Intel will put out 1.8As soon, maybe they won’t. I suspect these C1+s will turn out a bit better, but who knows when we’ll see them?

I don’t expect any real price drops until mid-January at earliest, and it may be more like three months (the next scheduled Intel cut is at the end of February).

Hibernating for the (Northern Hemisphere) winter may be the best bet.


Be the first to comment

Leave a Reply