Intel Engineer Speaks
Over in this forum, we have a person who claims to design customer
reference motherboards for Intel, and he’s working on Tualatin designs. If he’s faking it, he should win an award. 🙂
First the bad news:
You’re not going to be able to slip a Tualatin into your current setup (which is what most of you seem to want in order to buy it). There are pin incompatibilities, for one thing. While “Keith” thinks a slotket is theoretically possible, he is quite skeptical about getting one out at a reasonable price.
If you are in the market to buy a new Intel board, you will be completely out of your mind if you don’t wait for mobos that will support Tualatin. They should be coming out in the next few weeks. The Abit ST6 will support Tualatin, so will the Asus TUSL2, and there will be plenty of others.
The whole package is going to cost you, though, expect to spring over four Franklins for CPU and mobo sometime in July. Even more for the 512K version.
The good news is that the silicon runs pretty well. “Keith” has an unlocked early-silicon Tualatin doing 1.4Ghz with no problem.
However, we have two overclocking issues with Tualatin:
- How far can it go?
- How far can you go with it?
The answers can be quite different.
A few days ago, I predicted a max speed of about 1.4Ghz for this chip, based on past Intel historical practices. It is possible that Intel may be
doing something a bit different this time around and holding back on Tualatin speeds if they find that it whips a Willy. We’ll have to see.
However, you have that locked clock multiplier to deal with. The highest multiplier Tualatin will initially have is at most 9.5X. So even assuming the chip
is perfectly capable of sky-high speeds, we’re talking 1.425Ghz at 150Mhz, 1.52Ghz at 160Mhz, and a little better than that if Misters PCI or cheap memory don’t stop you.
A smaller piece of good news is that the server version of Tualatin (with 512K cache) will work fine on Tualatin desktop mobos. This could be very appealing to some.
Conclusions
TBird/Palomino vs Tualatin is going to turn into a PC/Mac argument. The first will holler “Cheaper, Faster!!!” while the other side will yell, “Solid, Reliable.”
I thought and think there is much wishful thinking about Tualatin being the perfect PIII upgrade chip, but wanting something doesn’t make it so.
I suspect that if we ever see a Tualatin solution for older systems; it won’t be a $15 slotket, but rather something much more involved and expensive and delayed.
Abit revises certain BIOSes for Coppermine-T
If you look at Abit’s Taiwan website, you’ll see a few fairly recent BIOS revisions for a few Intel mobos (BE6-II Rev 1.2, BX133, VH6 and VH6-II) with this little note:
Support Intel Coppermine T CPU.
Now just what is this? After seeing this, and poking around Intel’s site for a bit, my suspicion is the “Coppermine-T” is just a cD0 stepping Coppermine, it’s just a minor change, and here’s why:
- cD0 stepping Coppermines have a new CPUID (068Ah). A new CPUID usually calls for new microcode. There were no earlier BIOS revisions which would have covered the cD0 stepping.
- Intel is cheap. Why come up with one new stepping today, and yet something else in a month?
- There’s no real room on the roadmap for anything else. Tualatins will cover everything from 1.13Ghz up. PIII cD0s are announced up to 1Ghz. What ground is Coppermine-T supposed to cover?
If you’re still hoping that Coppermine-T is some sort of modified Tualatin that will let you run at around 1.5GHz or so on current equipment, very easy way to determine whether that’s the case of not. Any such beast is going to require its own VRM. No VRM circuitry, no modified Tualatin.
Sorry, but it looks like Intel wants you to buy a new mobo.
Be the first to comment