ATI released a statement to a few websites that with an additional conversation with the Inquirer about this.

While the statements are neither definitive nor entirely convincing, they (combined with other statements) are sufficient to change the light on the stop sign from red to yellow.

First, ATI repeats again and again that they’ve received few complaints. This is fairly meaningless since there aren’t that many AGP 8X boards out there, and of those sold, relatively few people would buy a $400 video card to go along with it.

Unfortunately, one spokesman shot himself in the foot by saving, “There have been very few AGP 8X problems reported, and many are due to issues unrelated to the graphics card itself.”

Many? Not all? If many aren’t due to the graphics card, that means some are.

The statement says the following, “From these tests, it has been concluded that no compatibility issues exist with the current or soon to be released AGP 8X motherboards.”

That would seem to be definitive, but in fact, it is not. It really boils down to what you call a “compatibility issue.”

Problem Number One

The first problem some people face is that their machine won’t boot up at all. Apparently, the people at Asus have described the problem here (entry dated 10:33 AM, September 16, 2002).

It looks like they’re saying that the ID on the R9700 sends incorrect information to the motherboard on initial powering up from a cold boot, and that a slight delay to allow a second read of the ASIC ID (either through the BIOS or through hardware) will solve the problem.

Sorry, ATI, this is a compatibility issue. Whether or not the issue is your fault is another question.

Whose “fault” is this? I’m sure ATI will say it’s the mobo manufacturer’s fault, and the mobo manufacturers will say that it’s ATI’s fault, and both of them can come up with a reasonable case. At this point, it probably would be safest to say that it’s both their faults.

There’s yet a third potential suspect in all this: the power supply. It is unclear whether or not this problem could get “fixed” by more power.

As you can see, this gets very complicated quickly. Nor can you cut through all this and place “fault” with whomever fixes it. If you mess something up, and somebody later makes adjustments to fix it, you would hardly automatically blame the fixer-uppers’ fault.

The important point to realize is that for at least the boot issue, this looks fixable.

This looks to be a problem that can be fixed by either or both parties, and probably is in the process of being fixed by both parties.

Problem Number Two

Once you manage to boot up, many are reporting 3D lockups.

This problem is going to be tough to sort out.

Some of it (likely a lot of it) may be due simply to overzealous overclocking.

There’s some indication that some people are finding that their DDR333 or 400 boards don’t, even when they have appropriate RAM. Some have reported that they’ve had to back down to 133MHz to get it to work.

Could that be because the boards su . . . (sorry, I’ll be PC) are FSB-challenged, and they’re doing the best that they can so it’s unfair to criticize them. 🙂

Or might it be that the R9700 demands power that the PS or mobo can’t deliver, and that is what is making some systems crumple?

Don’t know, but for the people fighting this, power is becoming a bigger suspect.

For those facing this problem, here’s a few tests:

1) Make sure you keep the AGP speed as close to 66MHz as you can. If this fixes the problem, then we probably have an overclocked AGP issue.

2) If that doesn’t work, the next good test would be to run one of these mobos with an earlier video card and see if the boards suddenly become less-FSB challenged. If they do, that would tend to point to the R9700 power.

3) Heat is a big factor with this card. It probably wouldn’t be a bad idea to use some temporary expedients to cool the card a bit more to see if that makes any difference.


At this point, it would probably still be best to avoid the R9700/AGP 8X combination for a while longer if you don’t want to do a lot of homework. Different manufacturers have or haven’t made changes, and even for those who have (for instance, Asus), it probably would be best to let those who have the boards work with them a bit more to make sure Problem Number One has been solved.

Outside of that, the red light has gone to yellow.

While Problem Number Two is more likely to consists of “normal” problems, it would probably be best to keep a wary eye on the 8X mobos until they’ve proven themselves more.

For everyone else, those of you for whom buying a bigger power supply would be a very unpleasant surprise might want to wait a bit to see if that isn’t an unofficial “requirement” for this card. I’m going to see if I can make some sense of that over the next few days.

Heat is an issue, period. If you expect to do some very heavy gaming, especially with an overclocked card, you’d better also expect to provide some additional cooling, too.

In a sentence, the R9700 looks like a high-priced, high-performance, high-maintenance card at the moment. Judge accordingly.


Be the first to comment

Leave a Reply