most stable current chipset for AMD?

[Tony Hill]

Was just wondering which would be the best chipset if one wanted to
build a pc today with an athlon xp.

nVidia nForce2 for sure, no question in my mind. VIA chipsets work,
but they tend to have just strange odd-ball problems, particularly if
you deviate from bog-standard add-in cards in the least. SiS chipsets
are fairly good, but their biggest problem is that they are put on the
cheapest hunk-of-crap motherboards, so the end result is the same as
if the chipset were shit. I also had somewhat more issues with my
previous SiS based system (admittedly a few years back) as compared to
my current nVidia system (which has been rather rock solid).

Other than that there is just ALi and AMD themselves. The AMD
chipsets are VERY dated and not really available anymore (AMD has
moved on to their Opteron/Athlon64 chipsets), and ALi has been.. umm..
kind of missing for the past few years, I don't know that anyone is
selling any boards with their chipsets anymore.

So, long story short, nVidia all the way IMO.

 

[Slug]

I couldn't find a SiS735-based motherboard at a decent price at a local
retailer, and wanted to avoid the NVidia chipsets,

And what is your logic behind wanting to avoid Nvidia chipsets?

 

[Scott Alfter]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

well for several months for digital content creation i've been running
an Athlon64 with VIA Albatron board, K8X800 proII - no problems so far
and yes this is my first VIA but i read the reviews that AMD64 boards
with VIA appeared to be much better than the previous generation...
here's hoping...

The SATA controller in VIA's Athlon 64 chipset has a problem that kept
corrupting the hard drive in a system at work. I worked around that by
switching to the Promise SATA controller that was also on the motherboard
(knock on wood ever since), but the Athlon 64 boxen I've built around
nVidia-chipset motherboards that were also equipped with SATA hard drives
have not had any storage-related problems (don't remember offhand if the
nForce3 supports SATA natively or if boards using it used an add-on
controller).

_/_
/ v \ Scott Alfter (remove the obvious to send mail)
(IIGS( http://alfter.us/ Top-posting!
\_^_/ rm -rf /bin/laden >What's the most annoying thing on Usenet?

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Linux)

iD8DBQFAsDcKVgTKos01OwkRAp5+AJ4vWN1if+NxtaOsMya1KnaflkrAPwCeJ9kN
Y3qjXEMnM3Pqt59k1E6naRc=
=3faS
-----END PGP SIGNATURE-----

 

[Ed]

Was just wondering which would be the best chipset if one wanted to
build a pc today with an athlon xp.

I'll vote Nforce 2. I've had an Asus A7N8X v2.00 (Nforce 2 Ultra-400
chipset) with Barton running WinXP Pro for just over a year now and it's
been flawless, running old bios and drivers too, no fuss, it just works!

Good Luck,
Ed

 

[GSV Three Minds in a Can]

I'll vote Nforce 2. I've had an Asus A7N8X v2.00 (Nforce 2 Ultra-400
chipset) with Barton running WinXP Pro for just over a year now and it's
been flawless, running old bios and drivers too, no fuss, it just works!

Yes, the nForce2 chipset is super. The current nForce3 chipset for AMD64
doesn't seem quite as slick, but maybe the revised one will be.

Having said that my nForce2 board (EPOX 8RDA+) went belly up yesterday
while recording some audio from cassette .. however it appears to be
'bad capacitors' rather than anything nVidia did.

(Anyone know when those dodgy-recipe electrolytic capacitors finally got
flushed from the build chain? I thought they were gone by early 2002,
but these boards were built in late 4q 2002, and seem to have them in
droves).

 

[Ed]

(Anyone know when those dodgy-recipe electrolytic capacitors finally got
flushed from the build chain? I thought they were gone by early 2002,
but these boards were built in late 4q 2002, and seem to have them in
droves).

What about the Asus K8x?? board, those were released not too long ago
with bad caps!
Ed

 

[gimp]

The SATA controller in VIA's Athlon 64 chipset has a problem that kept
corrupting the hard drive in a system at work. I worked around that by
switching to the Promise SATA controller that was also on the motherboard
(knock on wood ever since), but the Athlon 64 boxen I've built around
nVidia-chipset motherboards that were also equipped with SATA hard drives
have not had any storage-related problems (don't remember offhand if the
nForce3 supports SATA natively or if boards using it used an add-on
controller).

sorry to hear that :/ are u sure it wasn't a driver u installed or
didn't install...? i'm using onboard VIA SATA/RAID controller with a
raptor boot drive... no problems.

 

[Slug]

On Sun, 23 May 2004 05:30:50 GMT,

The SATA controller in VIA's Athlon 64 chipset has a problem that kept
corrupting the hard drive in a system at work. I worked around that by
switching to the Promise SATA controller that was also on the motherboard
(knock on wood ever since), but the Athlon 64 boxen I've built around
nVidia-chipset motherboards that were also equipped with SATA hard drives
have not had any storage-related problems (don't remember offhand if the
nForce3 supports SATA natively or if boards using it used an add-on
controller).

I used to have an AMD761 chipset mb and the only probelm with the mb
was the VIA usb controller they put on it. I had to do like you and
put in a PCI USB card to fix it.

 

[Scott Alfter]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

sorry to hear that :/ are u sure it wasn't a driver u installed or
didn't install...?

I used the driver that shipped with the motherboard. It might work better
with an updated driver, but it's working just fine through the Promise
controller. If it ain't broke...

_/_
/ v \ Scott Alfter (remove the obvious to send mail)
(IIGS( http://alfter.us/ Top-posting!
\_^_/ rm -rf /bin/laden >What's the most annoying thing on Usenet?

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Linux)

iD8DBQFAsY/+VgTKos01OwkRAsGdAJ9ldjOMb3EmsYTn1/5w5J/+n9WugwCg6nfr
w9573scJQZUfghN/i/66B3c=
=f4wY
-----END PGP SIGNATURE-----

 

[George Macdonald]

Just a lot of driver issues. This may have also been because back then I was
using Windows 98 and earlier on those systems, where drivers are just so
finicky to install and uninstall. If you ever change your processor or
motherboard to any other brand after you've had a VIA, you're going to have
a wonderful time with getting your OS back in working order again.

Before Win98, Intel had driver issues with i430HX et.al. as well as some
versions of the North Bridge which were broken. Surely you remember the
Win95 Bus Mastering driver of the month saga, which initially had no
uninstall procedure and you had to go through all the rigmarole of removing
..inf files, drivers and devices from device manager as each succeeding
version came out. Of course if you upgraded to Win98 without uninstalling
the Intel drivers.....BOOM!

No, apart from their reluctance to issue Errata Sheets, VIA's main problem
was that hardware add-in vendors, sound and other cards, CD/R-RW drives did
not validate their hardware with the VIA chipset.

With Win9x I've gone through several generations of chipsets, including to
and from VIA without a reinstall, the last from a VIA VP2 to a i440BX. You
just have to know how to do it and how to cope with any glitches along the
way.

But I have installed VIA based systems since those days for friends, but I
usually tell them strongly to avoid them if at all possible. If it's a
matter of cost, then I can't argue with them, VIA boards do often tend to be
cheaper than Nvidia, and more available than SiS.

For an Athlon64 system there isn't much else around other than VIA...
unless you want to use the nForce3 150 HT hobbled chipset. In fact some
mfrs, like MSI, have just avoided making nForce 150 mbrds. I've just done
a VIA K8T800 system because I got tired of waiting for nForce3 250 *and* I
expect it to be positioned at the high $$ end anyway.

This was a MSI K8T Neo mbrd and WinXP install went very smoothly and no
problems observed so far. VIA should look into this 4-in-1 thing though
and clean up the IDE filter/miniport driver confusion. One minor annoyance
with the system which is not unique to VIA, AFAICT, is that the HDD LED
does not work with a SATA drive. I'm not sure how much this is a hardware
problem with the mbrd - it appears that with *some* mfrs' mbrds a BIOS
update can fix it, while others need a new rev of the mbrd.

Rgds, George Macdonald

"Just because they're paranoid doesn't mean you're not psychotic" - Who, me??

 

[George Macdonald]

Was just wondering which would be the best chipset if one wanted to
build a pc today with an athlon xp.

I concur with the votes for nForce2. If you want to run DDR400 memory
choose it carefully though and look at the mbrd site for "approved" modules
- maybe check out some of the overclocking forums for experiences. Also
consider seriously if you might want to add memory later and what size to
start with and whether to go single or double sided DIMM - download any
mbrd manual and check details of supported configs.

Rgds, George Macdonald

"Just because they're paranoid doesn't mean you're not psychotic" - Who, me??

 

[George Macdonald]

Yes, the nForce2 chipset is super. The current nForce3 chipset for AMD64
doesn't seem quite as slick, but maybe the revised one will be.

Having said that my nForce2 board (EPOX 8RDA+) went belly up yesterday
while recording some audio from cassette .. however it appears to be
'bad capacitors' rather than anything nVidia did.

I dunno whether I'm being reasonable or not but I'm crossing Epox off my
list, as I had a 5-year old Epox mbrd suddenly go bad - dunno what it is
but it'd gone flakey and when I tried to run memtest86 on it, it died
completely.

(Anyone know when those dodgy-recipe electrolytic capacitors finally got
flushed from the build chain? I thought they were gone by early 2002,
but these boards were built in late 4q 2002, and seem to have them in
droves).

The IEEE article doesn't mention dates specifically and is kinda vague in
general about aportioning blame for obvious reasons:
http://www.spectrum.ieee.org/WEBONLY/resource/feb03/ncap.html

According to Gary Headlee, http://www.motherboardrepair.com/ it *does*
appear that 2001 was the cut off.

Rgds, George Macdonald

"Just because they're paranoid doesn't mean you're not psychotic" - Who, me??

 

[George Macdonald]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



I used the driver that shipped with the motherboard. It might work better
with an updated driver, but it's working just fine through the Promise
controller. If it ain't broke...

Which driver for which OS? With my MSI K8T Neo mbrd, for WinXP there's a
driver on a floppy to get the install going and there was a driver install
procedure on the mbrd CD, which I installed once XP was up and running.
It's early days but I haven't seen any problems yet. Blaming VIA's SATA
controller as the cause of your problem, without elaborating, seems like a
bit of a leap here... "Oh it must be VIA again".

Rgds, George Macdonald

"Just because they're paranoid doesn't mean you're not psychotic" - Who, me??

 

[Stacey]

George Macdonald wrote:

No, apart from their reluctance to issue Errata Sheets, VIA's main problem
was that hardware add-in vendors, sound and other cards, CD/R-RW drives
did not validate their hardware with the VIA chipset.

LOL!!

 

[nobody]

....snip...

The SATA controller in VIA's Athlon 64 chipset has a problem that kept
corrupting the hard drive in a system at work. I worked around that by
switching to the Promise SATA controller that was also on the motherboard
(knock on wood ever since), but the Athlon 64 boxen I've built around
nVidia-chipset motherboards that were also equipped with SATA hard drives
have not had any storage-related problems (don't remember offhand if the
nForce3 supports SATA natively or if boards using it used an add-on
controller).

Just recently set up MSI K8T Master2-far with 2xOpteron242 and SATA
drive (Hitachi Deskstar 160 MB) as a boot drive working off built-in
SATA controller. The only problem was that during WinXP setup I had
to feed the driver from a floppy (arrgh - who would think this
dinosaur-old technology may be still needed for anything!), but so far
no other problem noticed (though the system is only a few days old, so
I knock on wood). Sandra shows the performance in line with other
Opty systems, no matter AMD or VIA chipset, though such synthetic
benchmarks should be taken with quite a grain of salt.

 

[Tony Hill]

Before Win98, Intel had driver issues with i430HX et.al. as well as some
versions of the North Bridge which were broken. Surely you remember the
Win95 Bus Mastering driver of the month saga, which initially had no
uninstall procedure and you had to go through all the rigmarole of removing
.inf files, drivers and devices from device manager as each succeeding
version came out. Of course if you upgraded to Win98 without uninstalling
the Intel drivers.....BOOM!

Ugg, now THAT was a nightmare! I was working at a computer store for
some of this time frame and we had to be SUPER careful with the order
we installed drivers or else the whole system was just screwed. Even
while being careful we sometimes wouldn't know what patches a customer
had installed previously and things just ended up being a huge mess.

However Intel DID eventually manage to get stable drivers out.
Sometimes it would take Intel 6-8 months, but they would eventually
get things working. That, in my experience, is VIA's main problem,
they just never seem to quite fix the problems, only reduce the
symptoms.

No, apart from their reluctance to issue Errata Sheets, VIA's main problem
was that hardware add-in vendors, sound and other cards, CD/R-RW drives did
not validate their hardware with the VIA chipset.

That used to be the problem, and that's a large part of the reason why
things aren't as bad now as they used to be way back in they day.
However since about mid-'98 it just hasn't been an option not to
validate your hardware with VIA chipsets, there are just WAY too many
people using their chipsets to ignore them (unless you're Creative
Labs apparently). No, VIA's problems run deeper than that.

With Win9x I've gone through several generations of chipsets, including to
and from VIA without a reinstall, the last from a VIA VP2 to a i440BX. You
just have to know how to do it and how to cope with any glitches along the
way.

You mean glitches like having to use Win9x? That's a pretty big
"glitch" IMO :>

For an Athlon64 system there isn't much else around other than VIA...
unless you want to use the nForce3 150 HT hobbled chipset. In fact some
mfrs, like MSI, have just avoided making nForce 150 mbrds. I've just done
a VIA K8T800 system because I got tired of waiting for nForce3 250 *and* I
expect it to be positioned at the high $$ end anyway.

The nForce3 250 is *FINALLY* showing up on store shelves, albeit
slowly. You're pretty much hitting on the main reason I've been
reluctant to recommend any Athlon64 systems to people, bad
motherboards. VIA does seem to have improved a bit, though I'm still
not sure I trust them; I just ran into too many odd little glitches
that never seemed to get fixed. Usually nothing that could easily be
pinned as for sure being a chipset issue, but they just didn't happen
on non-VIA boards. Things like problems with my audio depending on
what was being displayed by the video card, or issues with the network
card and a different video card. These things just worked so much
more smoothly in an nVidia or Intel based board.

 

[Stacey]

VIA does seem to have improved a bit, though I'm still
not sure I trust them; I just ran into too many odd little glitches
that never seemed to get fixed. Usually nothing that could easily be
pinned as for sure being a chipset issue, but they just didn't happen
on non-VIA boards. Things like problems with my audio depending on
what was being displayed by the video card, or issues with the network
card and a different video card. These things just worked so much
more smoothly in an nVidia or Intel based board.

Like you said the "problems" could be sorted out. Like swapping out
different video card/sound card combos that would coexist with the via
drivers etc. The deal is with the other chipsets this didn't need to be
done. I look at it as "I save $10 buying a via board and end up spending 2
hours extra fighting with it." My time is worth a little more than $5 an
hour.

 

[George Macdonald]

Ugg, now THAT was a nightmare! I was working at a computer store for
some of this time frame and we had to be SUPER careful with the order
we installed drivers or else the whole system was just screwed. Even
while being careful we sometimes wouldn't know what patches a customer
had installed previously and things just ended up being a huge mess.

However Intel DID eventually manage to get stable drivers out.
Sometimes it would take Intel 6-8 months, but they would eventually
get things working. That, in my experience, is VIA's main problem,
they just never seem to quite fix the problems, only reduce the
symptoms.

Well Intel *did* manage to get M$ to do it "their way" for things like IDE
BM drivers. That made a big difference for end-users who saw that there
were no add-on drivers to get the OS working on an Intel chipset. It took
a bit longer for M$ to waken up on this and the pidgin English instructions
on the VIATech Web site didn't help back then and that hasn't changed much.

That used to be the problem, and that's a large part of the reason why
things aren't as bad now as they used to be way back in they day.
However since about mid-'98 it just hasn't been an option not to
validate your hardware with VIA chipsets, there are just WAY too many
people using their chipsets to ignore them (unless you're Creative
Labs apparently). No, VIA's problems run deeper than that.

Even in mid-'98 timeframe there was still the residue of the writeable
ATAPI interface foul-up. Of course CD-R/RW drive makers just did it "their
way" in the absence of a standard method and continued to do so.

The nForce3 250 is *FINALLY* showing up on store shelves, albeit
slowly. You're pretty much hitting on the main reason I've been
reluctant to recommend any Athlon64 systems to people, bad
motherboards. VIA does seem to have improved a bit, though I'm still
not sure I trust them; I just ran into too many odd little glitches
that never seemed to get fixed. Usually nothing that could easily be
pinned as for sure being a chipset issue, but they just didn't happen
on non-VIA boards. Things like problems with my audio depending on
what was being displayed by the video card, or issues with the network
card and a different video card. These things just worked so much
more smoothly in an nVidia or Intel based board.

Yeah but from what I see, initial mbrd mfrs' efforts on nForce3 250 are on
the Gb version... to get the extra $$?? Fine for a home system, where you
want bells 'n' whistles but bugger-all use for an office system.

No doubt there was a VIA problem as demonstrated by the "Promise Bus
Mastering" priority, abort, retry thing. OTOH, most of the VIA complaints
were noise from people who couldn't follow instructions even in perfect
English.

Rgds, George Macdonald

"Just because they're paranoid doesn't mean you're not psychotic" - Who, me??

 

[Ed O'c]

Kai Harrekilde-Petersen wrote:


I'm *not* in the business of supporting systems that I've built.
There is only so much I'm willing to do in exchange for dinner
and a few beers. I want to put the pieces together, install the
OS and drivers, hand it over to the user, and be done with it.
I can do that with an nForce2 motherboard. I can seldom do that
with a VIA motherboard.

Frankly, agree fully. Built a Via system and have had so many
compatibility problems with standard components in the market. I will
never use a VIA motherboards again. On the other hand, a SIS system I
built never had the same problems. Lucky I had the SIS system to
accept the many parts the Via system would have nothing to do with.

 

[neilslade]

I sent Gary Headlee my video card, which he also claims to repair--

a MONTH ago.

His web site states turn around time of 1-2 days.

I finally had to call the POLICE to track the guy down, after sending
him
cash and my card-- received by him via delivery confirmation.

He called me back, and promised to look at my card within a few days.

NOTHING- weeks later.

He does not answer EMAIL, and he has an UNLISTED PHONE NUMBER.

He still has my video card worth a couple hundred bucks.
He asks that people send him CASH for repairs.

You figure it out.


http://www.motherboardrepair.com/index.php?sec=home


Neil Slade Books and Music
Denver, CO 80206