There are always athlon boards that run a AMD chipset, like the
760MP/MPX.

 I've heard nothing but good about them.

 As a side note, I would not trust any of my data on a maxtor HDD. I've
personally seen and heard of far more failures on these drives than any
other put together.


itai at replic d0t net

On Sun, 23 Dec 2001, Peter Seebach wrote:

> In message <E16IJBR-000Kf5-0V@anchor-post-31.mail.demon.net>, Graham/Aniartia w
> rites:
> >> i815 based motherboard looks stable for me and fast .
>
> >Why go for an Intel CPU? They're expencive & slow.
>
> Because most of the AMD-friendly motherboards have VIA chipsets.
>
> What's wrong with those?  How shall I count the ways:
> * At BSDi, I spent a couple of calendar months (figure a total of 4+ working
>   days, solid) trying to debug a customer's problem with serial ports that
>   would, every so often, just stop delivering interrupts.  The problem?  VIA
>   chipset; he replaced the board with an Intel chipset-based board, and it
>   got better.
> * I have one motherboard that will, every couple of months, lose the ability
>   to see hard drives after a boot, until it's been powered down for at least
>   two hours.  VIA chipset.
> * I had (I returned it) a motherboard where NetBSD would, 100% reliably,
>   see the first key you pressed after boot as held down indefinitely until
>   you unplugged the keyboard and plugged it back in.  VIA chipset.
> * The Abit VP6 says it supports ECC, but it doesn't.  VIA chipset.
> * One of the motherboards I have used to start doing the "I have no valid
>   video card" thing after about three boots, until you removed the video
>   card and replaced it.  This did only happen with one specific model of
>   video card - but it also only happened with this motherboard.  VIA chipset.
> * I used to have a motherboard that couldn't even boot with a 5GB drive
>   installed, no matter what it was jumpered as, or how you set the BIOS up.
>   VIA chipset.
> * I have returned at least one or two other motherboards for humorous or
>   implausible failings - all with VIA chipsets.
>
> By contrast, I have had the following problems with Intel-based chipsets:
> * A machine was dropped 15', such that, after shearing off its mounting
>   screws, the drive cage nearly cut the SCSI cable in two.  A year or so
>   later, the board spent two hours plugged in, sans surge protection, to an
>   outlet that was wobbling between 0 and 60v AC, such that the power supply
>   made a very strange humming sound and a bit of a burning rubber spell.
>   After this, the board stopped working.
>
> Now, to be fair, the i810 and i815 chipsets don't support ECC - but they
> somehow seem to have avoided the problem where vendors go around claiming
> they do.
>
> I'm not saying that VIA chipsets are cheap pieces of crap; I'm just saying
> that every last single one I've ever had to deal with generally has been.
> I was convinced to try a recent Tyan board that has some AMD parts, and some
> VIA parts, and it looks like it's a lot better; they may only suck at some
> parts of this, or AMD may have whipped them at least partially into shape,
> but the sheer ubiquity of their insanely weird falure modes has me pretty
> much convinced that, if I'm going to try to build a reliable machine, I'm
> going to avoid things that are based on VIA chipsets.
>
> So, yes, I'll take slightly worse Intel CPU's if it gets me a motherboard
> chipset with decent performance.
>
> While we're at it, I'd like to take this opportunity to wish devoutly that
> I could find a board with a ServerWorks chipset that did Athlons.  I have used
> a number of ServerWorks chipsets, and indeed, their HE-SL chipset just sold me
> a couple of computers - because the computers *I* built for trying to run
> gigabit networking could actually saturate a local area gigabit LAN, and
> someone else's couldn't.  :)  (I'm quite happy to say that apparently the
> current record for *WAN* connections with those boxes is a bit over 600Mb/s.)
>
> -s
>