Is the KX133 broken? Read this!

Hardware

Golden Member
Oct 9, 1999
1,580
0
0
Could GHz Athlon glitch signal deeper problems?
by Scott "Damage" Wasson - 01:00 am, July 5, 2000

Everybody is finally reporting on the problem Gateway's found with 1GHz Athlon T-bird systems (they lock up). I was waiting to hear more detail on this one before posting, since I'd seen something about this a while ago along with mention that the issue had been resolved. Not really sure what the story is here, but odds are it has something to do with A) the motherboards Gateway has chosen to use, B) Via's KT133 chipset, or C) some combination of those two things. The problem doesn't seem to affect all 1GHz T-bird systems.
My best guess: Via's AGP bus implementation is just wrong somehow. So far, I've used two boards based on the very-much related KX133 chipset?though neither with T-birds?and I've seen lock-ups in 3D apps with both of them. The problem seems to be related to the CPU clock frequency, but it's independent of overclocking. I've "tested" this problem more extensively that I'd care to admit trying to make things work right on a couple of systems here, and I've used both mildly overclocked CPUs and an 800MHz Athlon at its stock speed (and below). AGP-related lock-ups occur at CPU frequencies above about 700MHz. Turning off AGP 4X, fast writes, etc. helps, but it's no cure.

Then there's the still slightly mysterious problem with the KX133 chipset that caused AMD to refrain from making Slot A T-birds widely available via retail. Speculation about the nature of the KX133/T-bird problem has centered on the way the CPU signaling is handled, but I tend to think there's more to it. Maybe the KT133 is better?it may handle CPU signaling right?but I don't think it's only problem. The "fixed" KT133 seems to be having AGP lock-up problems, as well.

The long and the short of it: perhaps neither of Via's Athlon chipsets is right.

Anyhow, that's my take. Maybe it's just frustration talking, but I've tested enough different ways and isolated enough different variables?yanking out sound cards, swapping video cards, installing newer/older drivers?to be convinced. For now, I'd recommend an AMD 750-based mobo or (egads!) an Intel-based system instead of any Athlon/Via combo.


...........................................
This is exact my experience! DAMM VIA
 

Ulysses

Platinum Member
Jun 17, 2000
2,136
0
0
Of course the KX133 chipset is not 'approved' by AMD for the Tbird. Only the Irongate 750/751 and the newer KT133 are, if I'm not mistaken. Would Gateway, or any OEM, then be using the the non-spec KX133 in any of their systems - especially in premium 1 GHz sytems?

OTOH, AMD may work with OEM's and have different standards for their systems. I assumed that even though AMD has made Slot A new Athlons available to OEM's, those CPU's were being installed in KT133 boards. Maybe that's wrong.

I'm not that influenced by reports by individuals of problems, or success, with their home-built systems. If the big OEM's use a chipset that tends to indicate it's probably OK (except for the exception that proves this rule, the Intel 820 (!!!) - but even the 820's problems were rather arcane, specific and limited). The repurcussions for OEM's of problems with their systems can be great - like for Packard Bell. It takes extensive and highly technical testing to show the chinks in a chipset or system, I don't think just an individual 'swapping' cards anb stuff proves much. But I'm not trying to defend VIA chipsets either.

Apparently some KX133 boards may appear to be 'compatible' with Tbird and some may not. See:
http://www.vr-zone.com/

We may never know why Gateway is having trouble with the 1 GHz sytems. Compaq had some 'out of memory' problems reported recently with some of their better consumer systems, and I never heard what the final cause was (probably too many 'helpful' utilities put in the system tray at the factory).
:)