Old news - VIA KT400A confirmed

CougTek

Hairy Aussie
Joined
Jan 21, 2002
Messages
8,729
Location
Québec, Québec
Seen on OCWorkbench :
After chatting with a few manufacturers, it is confirmed that most of the top tier will come out with a model for the VIA KT400A chipset. As you probably already know, KT400A does not have significant peformance difference frmo the KT400 chipset. The only thing that is probably different is the support of DDR400 and probably native support of some newer features using a newer south bridge.

ASUS will be relasing A7V400 in Mid-March featuring this new chipset.
I'm sure you are all very excited after having read this.

nForce 2 and SiS746FX are more interesting IMO.
 

blakerwry

Storage? I am Storage!
Joined
Oct 12, 2002
Messages
4,203
Location
Kansas City, USA
Website
justblake.com
Since the nforce was released and gained popularity I put VIA on the backburner, and reserved them for "budget" boards... meaning I'd probably get a kt266(a) and forget the 333 and 400 even existed.

The kt333 and kt400 really offer little benefit over the kt266a, however boards based on them sometimes cost double that of comparable kt266 boards.
 

Mercutio

Fatwah on Western Digital
Joined
Jan 17, 2002
Messages
22,275
Location
I am omnipresent
Yeah. Inexpensive stability sucks. At the $5 a KT333 with USB2 support costs over the price of KT266A? That's not worth a damn thing!

nforce2 is a niche board. The high-end audio is cool, but mostly I don't see it as delivering real value to a customer. If I tell someone they can pay $10 more and get USB2 and firewire (lots of KT400s have these), that's something they get excited about. If I try to explain "Dual Channel DDR" or "support for high FSBs you aren't ever gonna use", we're into eyes-glazing-over territory.
 

blakerwry

Storage? I am Storage!
Joined
Oct 12, 2002
Messages
4,203
Location
Kansas City, USA
Website
justblake.com
i could get the Epox EP-8K3AE for $70.... but no usb 2.0...

why not just buy the Shuttle AK32L/A for $53?

if you're gonna buy a board with no features, it might as well be a shuttle... :eekers:



The Nforce2 offers all the features of kt400, with a few benefits. 1) Cost of a full featured board is less 2) integrated firewire 3) better sound 4) better stability record 5) no 4-in-1 6) instant performance increase w/ possible more performance using dual channel DDR.

I remember just a few months ago Soyo's top of the line kt333 was ~$130... their kt400 even higher... now a comparable board, based on the nForce2, the Asus a7n8x is less. And the Asus deluxe model is clearly superior with S-ATA, dual NIC, and superior sound.
 

blakerwry

Storage? I am Storage!
Joined
Oct 12, 2002
Messages
4,203
Location
Kansas City, USA
Website
justblake.com
Xbit labs kt400 vs nForce2 said:
Conclusion
Now let's try to sum it all up. The results we've got during the tests indicate that the situation in the Socket A chipset market may change soon. If nothing extraordinary happens, we may see the fall of the Taiwanese VIA in this field. VIA's got a strong and resourceful rival, NVIDIA, that has learnt from its technological and marketing mistakes and is ready to fight for the customer.

Our investigation suggests that the nForce2 chipset from NVIDIA is a high-performing and multifunctional solution for advanced users.

To be more specific, nForce2 works best with CPUs that have 333MHz bus and that's exactly the product to be offered by AMD in 2003. So, the combination of nForce2 and Barton-based Athlon XP will probably be the number one choice as a fast platform for AMD fans.

VIA's products, KT333 and KT400, have only one future: they will be used in low-end systems with 266MHz bus CPUs. We should definitely admit that KT333 and KT400 provide best performance with these CPUs.

Well, it is also quite possible that VIA will roll out one more version of its chipset, VIA KT400A, optimized for CPUs with 333MHz bus. If it happens, VIA will have some share in the high performance systems field, although its chipsets have not very impressive specs compared to NVIDIA nForce2.

If we compare VIA KT400 and VIA KT333, we should mention that the first one does perform slightly faster with 266MHz bus. But with 333MHz bus, both chipsets from VIA perform about the same.

As for the memory, the situation looks as follows. DDR400 SDRAM that costs more than DDR266 and DDR333 provides no performance growth no matter how high its theoretical bandwidth might be. Moreover, DDR400 often slows down Socket A systems. The most appropriate choice for Socket A platforms today is DDR333 SDRAM. It provides highest performance in any modern system. The only exception is nForce2 that is heavily optimized for synchronous work of the memory and CPU buses. So, with 266MHz system bus, you'd better use DDR266 SDRAM in an nForce2 based system.
 

Mercutio

Fatwah on Western Digital
Joined
Jan 17, 2002
Messages
22,275
Location
I am omnipresent
I have couple of problems with nforce. One is that it's made by nvidia, a company whose products I firmly - and with good reason - believe are inferior. The second is that as a chipset, it hasn't been widely adopted by companies whose products I would actually buy (i.e. Gigabyte doesn't do it, Tyan doesn't do it. Epox does, which to me is a good sign, but Asus, MSI and Abit are collectively worth a bucket of warm spit IMO). I won't even dignify comments about nvidia stability vs. VIA with a response. Suffice to say I'd have to take issue with the suggestion that nvidia's is better somehow.

I guess if there was a "part 3", it's that there aren't a whole hell of a lot of 333FSB Athlons out there yet. I pretty much push CPUs in the $70 to $100 range. I guess the 2400+ could make it there in the next 6 months, but looking at current pricing, I wouldn't hold my breath.

Performance? Who cares? I stopped looking at performance numbers for motherboards in 1998. nforce is 3% faster than KT400. I'll alert the media.

I buy Gigabyte KT400 boards. They're utterly fuss-free, completely stable and have the features I want to provide to customers. As far as I'm concerned, I'm paying for quality.
 

blakerwry

Storage? I am Storage!
Joined
Oct 12, 2002
Messages
4,203
Location
Kansas City, USA
Website
justblake.com
I guess we are shopping for two different kinds of people. I look at the Asus a7n8x and think "wow, that motherboard is nice, I'd like to upgrade to that"

so when someone asks my opinion on a mobo, I recomend the Asus a7n8x...

If they want a cheaper mobo, I recomend the Epox nForce2 board... if they want an even cheaper board, i recomend the shuttle AK32A...

If they want intel, I simply recomend the 845GE.
 

time

Storage? I am Storage!
Joined
Jan 18, 2002
Messages
4,932
Location
Brisbane, Oz
Blakerwry said:
The kt333 and kt400 really offer little benefit over the kt266a, however boards based on them sometimes cost double that of comparable kt266 boards.
WTF? Support for USB 2.0, PC2700, 333MHz FSB, ATA133, improved PCI, double the North-South bandwidth, is all little benefit? And where the hell did you see them for twice the price?

The Nforce2 offers all the features of kt400, with a few benefits.
1) Cost of a full featured board is less
Wrong.
3) better sound
Most NForce2 boards are using the same third party audio chips as KT400.
4) better stability record
Show me some evidence.
5) no 4-in-1
It's called "Unified Driver Package" instead.
6) instant performance increase w/ possible more performance using dual channel DDR.
Firstly, with the majority of applications, performance is indistinguishable from a decent KT400 implementation such as Gigabyte 7VAX - in fact the latter is sometimes faster. Adding the second DDR channel improves performance from between zip to 3% - inside the margin of error for most benchmarks and absolutely unnoticeable.
 

Tea

Storage? I am Storage!
Joined
Jan 15, 2002
Messages
3,749
Location
27a No Fixed Address, Oz.
Website
www.redhill.net.au
It's quite astonishing to me that zome people are zo lacking in the ability to diztinguish between solid, dependable performance and marketing fud that they ztill think that VIA chipsets are zomehow "inferior". I have no idea who otherwise intelligent people continue to peddle this ztupid myth. If I want zomething that is a little different, zure, I'll buy a Nvidia or a SiS or zomething. But if I just want something that woks and I don't have to worry about, it'z VIA every time.
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
Tea said:
It's quite astonishing to me that zome people are zo lacking in the ability to diztinguish between solid, dependable performance and marketing fud that they ztill think that VIA chipsets are zomehow "inferior". I have no idea who otherwise intelligent people continue to peddle this ztupid myth. If I want zomething that is a little different, zure, I'll buy a Nvidia or a SiS or zomething. But if I just want something that woks and I don't have to worry about, it'z VIA every time.
You mizzed a couple Tea.
 

Tea

Storage? I am Storage!
Joined
Jan 15, 2002
Messages
3,749
Location
27a No Fixed Address, Oz.
Website
www.redhill.net.au
Zorry.

Blame Tannin. He iz a bad influence on me zometimez. Alzo, thiz hot weather haz the effect of making my brain go all zoft and gooey and dribbling out my earz if I am not careful. Zometimez I feel almost ztupid enough that if I zhaved my fur off and wore a zuit people would miztake me for a human.

(No dizrezpect to my wonderful human friendz intended, of courze.)
 

CougTek

Hairy Aussie
Joined
Jan 21, 2002
Messages
8,729
Location
Québec, Québec
The performance difference between the KT400 and the nForce2 isn't at default 133MHz FSB (where they are both relatively equal), it's at 166MHz FSB. The nForce2 with its memory slots populated correctly, is ~5-10% faster than the KT400 at this FSB frequency, which is starting to be pretty significant just for a chipset difference. In most popular games and memory-heavy applications at least. If all you intend to do is MS Office and net-browsing, of course a nForce2 makes little sense.

If you don't plan to overclock your cheap Thoroughbred B to 166MHz FSB and if you don't need the features of the nForce2 chipset (the integrated APU audio is the only real unique feature, as integrated LAN is common to almost every motherboard out there), then save some bucks and go for the KT333CF or KT400.

However, currently, there are no better bang for the buck than a lower-end T'bred working at 166MHz FSB on a nForce2-based motherboard with the MCP-T south bridge, at least for the enthousiasts/gamers audience.

And one last thing :
time said:
Most NForce2 boards are using the same third party audio chips as KT400.
They share the same audio codec (Realtek ALC650), but the difference is that one doesn't use the host processor to perform all the calculation while the other (KTxxx) do. The end-result is that the APU of the nForce2 saves you several CPU cycles for the other stuff you're doing meanwhile. This also means that you'll heard less "pop" every now and then on a nForce2 than on a KTxxx when listening music while the CPU is busy, since the audio calculation won't depend nearly as much on the host processor. So while it's true that both share the same codec, it isn't true that their audio quality is comparable. It isn't.
 

time

Storage? I am Storage!
Joined
Jan 18, 2002
Messages
4,932
Location
Brisbane, Oz
Thanks for the clarifications, Coug.

Just for the record, I didn't mean to appear to be bashing the nForce 2 - I've already got one here. It's just that I can easily see Mercutio's point about the board package being more important to many people.

The vast majority of people will never overclock their CPU, and not many really need an extra 10-15% in games (that's 4-6fps at 40fps). Like most people here, the idea of 'free' extra performance appeals to me, but I couldn't help noticing that the GA-7VAX is closer than I'd been led to believe. And it's definitely a professionally designed package that I can sell with confidence.

On the other hand, nForce boards like the Epox finally offer Firewire. And with CPU speed increments as little as 4%, 10% could represent two or three speed grades.
 

Newtun

Storage is nice, especially if it doesn't rotate
Joined
Nov 21, 2002
Messages
485
Location
Virginia
CougTek said:
And one last thing :
time said:
Most NForce2 boards are using the same third party audio chips as KT400.
They share the same audio codec (Realtek ALC650), but the difference is that one doesn't use the host processor to perform all the calculation while the other (KTxxx) do. The end-result is that the APU of the nForce2 saves you several CPU cycles for the other stuff you're doing meanwhile. This also means that you'll heard less "pop" every now and then on a nForce2 than on a KTxxx when listening music while the CPU is busy, since the audio calculation won't depend nearly as much on the host processor. So while it's true that both share the same codec, it isn't true that their audio quality is comparable. It isn't.
But in my infoquest for nF², I came across the nForce audio forum at nForcersHQ.com.
  • Top item: 3-page sticky on "crackling" audio
    [*]NForce2 No Audio
    [*]NFORCE2 AND USB SOUND PROBLEM!!!
    [*]Nforce Crackling [ Goto page: 1 . . . 19 ]
    [*]Line In problems for new shuttle *Help*
    [*]MY SOUND CHIP WENT IN FLAMES!!!!!!
    [*]A7N8X Standard onboard sound bad
    [*]A7N8X sound muddled?? Help!
    [*]can't install win98 "nvidia nforce audio" drivers
    [*]Major 8RDA+ Audio Problem!
    [*]Epox 8RA+ : no sound for some things
    [*]Soundproblems on Leadtek k7ncr18d pro
    [*]background Hiss in games?
    [*]Strange Sound Problem on A7N8X Deluxe
Reminds me of the VIA/SoundBlaster controversy.
 

blakerwry

Storage? I am Storage!
Joined
Oct 12, 2002
Messages
4,203
Location
Kansas City, USA
Website
justblake.com
how many billions of threads have been devoted to "AC'97 doesn't work"... "...skips" ..."sux0rs"


Coug is absolutely correct, the kt400 was designed for 133mhz operationm not 166mHz(even though it can run it)... 166mHz is what kt400a and nForce2 was designed for.

Personally, I feel that upgrading from a kt266a to a 333 or 400 was pointless. Sure, if you had an old BX board, you might as well get a kt400 if it wasn't much more expensive. The problem was up until the Nforce2 was about to be released It *was* more expensive.

For most things the kt400 offers a small increase in speed over the kt266a with the benefits of AGP3.0 and USB2... the kt333 offers almost no benefit over the kt266a however...


I agree that VIA based boards are good, and half my computers are VIA based... however, I just think that for a new "fully loaded" machine the nForce2 is a better choice because of its features, speed, and competitive price.
 

zx

Learning Storage Performance
Joined
Nov 22, 2002
Messages
287
Location
Beauport, Québec, Canada
Maybe I have to remind people of good old KT133A and KT266 (without the A). I had one board based on both, and it was slow, shitty and unstable. The board with the KT266 (MSI) actually never worked well. It had stability problems from the start that ranged from frequent lock-ups to very regular programs having errors (winamp, IE) to BSOD's. And yes, I tried installing reformatting many times. I underclocked the CPU and memory and nothing would do. When it was switched to an AOpen KT266A board, all problems vanished.

Same thing with the KT133A. At least is wasn't as bad as the KT266, but it really sucked. Corruption errors with files, BSOD's when using a cable modem on USB, very slow ATA (with abnormally high CPU utilisation) both for disks and for CD-ROMs (when ripping CDs or burning CD's, my computer was unresponsive until it finished).

Maybe VIA has improved though. I had a friend with a KT266A (gigabyte board) and it worked like a charm. Maybe the horrific 686B south bridge gave VIA a very bad name. I suppose that the same thing could happen to nVidia, and that would also give them a very bad reputation.

I never had an experience with a NVidia board though. However, i'm pretty sure that SiS is much better than VIA. They are very close to Intel's quality. It's a shame that there are so few socket A boards based on a SiS chipset.
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
zx said:
the KT133A. At least is wasn't as bad as the KT266, but it really sucked. Corruption errors with files, BSOD's when using a cable modem on USB, very slow ATA (with abnormally high CPU utilisation) both for disks and for CD-ROMs (when ripping CDs or burning CD's, my computer was unresponsive until it finished).
A lot of KT133A boards were plagued by such problems (Abit's in particular seemed to be problematic). Nonetheless, some weren't. For example (although I know that Tannin will disagree cause he had some sort of weird problem which I can't remember offhand, but for me:) MSI K7T-Turbo -- I have never had any problems. Its a very stable platform offering decent performance.

However, from all that I read, it does seem that the MSI KT266 offering was indeed a dog....saw so many people report problems with that board. I don't know if that typified the whole gamet of KT266 chipset based boards.

IMO:
KT266A - appeared to be a pretty good chipset (albeit, lingering PCI related defficiencies .... but, despite the uproar/backlash, the avg. user simply wouldn't notice them unless told about them).
KT333 - with the exception of (so I hear) a few boards that can offer sychronous 166/166 operation, boards based on this chipset offered little over its predecessor.
KT400 - some nice additions and capabilities, plus PCI should be resolved and related problems are manufacturer-BIOS specific.
KT400A - expect it to be a mature platform with solid performance and stability --- all the chinks in Via's armour should be hammered out....unfortunately for them, its a little late in the K7 game being that K8 Hammer is coming in the fall....as such, it will be interesting to see how well boards based on the 400A will do in the marketplace.

It's a shame that there are so few socket A boards based on a SiS chipset.
I concur.

CK
 

Tannin

Storage? I am Storage!
Joined
Jan 15, 2002
Messages
4,448
Location
Huon Valley, Tasmania
Website
www.redhill.net.au
CityK said:
I know that Tannin will disagree cause he had some sort of weird problem which I can't remember offhand, but for me:) MSI K7T-Turbo -- I have never had any problems.

Well, if you call having five boards and having to replace all five of them on RMA, one by one (or possibly it was 4 of them and one of them twice), and then having to replace three or four of the replacements as well, and having (I think) just one out of a total of about 12 or 14 boards supplied to us one way or another actually still working and in service "weird", I guess I had a weird problem with that particular MSI board. The retail boxed one was OK, we had a small number of those too, and no problem to speak of, but the allegedly "identical" OEM version was appallingly bad.

In fact, it was the only bad KT-133A I've ever seen or heard of. Like most VIA products, it was an excellent chipset and we sold many, many hundreds of them: Soltek, Epox, Gigabyte, and yes, even MSI in volume, plus various assorted others in small quantities, including an ASUS which was not fantastic but at least soilid - and ASUS in those days were capable of buggering up any chipset to make the board as difficult to use as possible. In CK's words, "It's a very stable platform offering decent performance."
 

zx

Learning Storage Performance
Joined
Nov 22, 2002
Messages
287
Location
Beauport, Québec, Canada
So you say that the bad reputation that VIA got was due to bad individual boards? How come this only happened to VIA?

I got a SiS chipset on an ECS board that works perfectly. It's much faster and stable than my old KT7A. I mean ECS isen't in the best in terms of quality... I did not see any chipset related problems with other socket A chipsets like the nForceX or the SIS7xx.
 

Tannin

Storage? I am Storage!
Joined
Jan 15, 2002
Messages
4,448
Location
Huon Valley, Tasmania
Website
www.redhill.net.au
VIA doesn't have a bad reputation (except among the ignorant), so no, I'm not saying that. Sure, there were bad boards, but that's probably true of every chipset ever made. The vast majority have been good, and many have been truly excellent.

And don't even get me started on those ultra-crap ECS SiS things! But then, ECS could make a bad board with a BX chipset - and they probably did!
 

zx

Learning Storage Performance
Joined
Nov 22, 2002
Messages
287
Location
Beauport, Québec, Canada
Why is :

SiS735 + ECS (cheap crappy shit) = good board

KT266 + MSI (mi-range?) = crappy board

KT133A + Abit(variable quality?) = crappy board

Tannin, have you tried other socket A chipsets?

Also, if you can't notice that the KT133A's ATA controller is bad when you do not burn CD's or copy CD's. With hard drives, it's only slower than other ATA controllers. With CD's the system is unresponsive.
 

time

Storage? I am Storage!
Joined
Jan 18, 2002
Messages
4,932
Location
Brisbane, Oz
zx said:
SiS735 + ECS (cheap crappy shit) = good board
That's a contentious statement. The last time this was discussed here, some people were happy, some were very upset - but I don't think anyone thought it was the best board around. I think inconsistency is the watchword with ECS, but surely you can see that that is inevitable from a discount manufacturer?

KT266 + MSI (mi-range?) = crappy board
Most definitely. However, people like me avoided all KT266, so maybe it wasn't all MSI's fault?

KT133A + Abit(variable quality?) = crappy board
Absolutely. If I had a dollar for every Abit KT7A owner whining about how bad Via is, I could buy out Bill. Mind you, I'm sure the majority of owners were happy, but there's no denying a vocal percentage had endless grief.

Also, if you can't notice that the KT133A's ATA controller is bad when you do not burn CD's or copy CD's. With hard drives, it's only slower than other ATA controllers. With CD's the system is unresponsive.
Total BS. Don't include all makers and boards in your claim just because you have a dud.

Much has been made of the limited PCI bandwidth available in KT133A solutions. This only affected maximum transfer rates on some RAID setups, and was competely irrelevant to almost everyone. Not to be confused with the SB issue, or course, but as I recall that was partly down to certain manufacturers (Abit) fiddling with settings that were reserved by Via.
 

zx

Learning Storage Performance
Joined
Nov 22, 2002
Messages
287
Location
Beauport, Québec, Canada
time said:
Also, if you can't notice that the KT133A's ATA controller is bad when you do not burn CD's or copy CD's. With hard drives, it's only slower than other ATA controllers. With CD's the system is unresponsive.
Total BS. Don't include all makers and boards in your claim just because you have a dud.

Much has been made of the limited PCI bandwidth available in KT133A solutions. This only affected maximum transfer rates on some RAID setups, and was competely irrelevant to almost everyone. Not to be confused with the SB issue, or course, but as I recall that was partly down to certain manufacturers (Abit) fiddling with settings that were reserved by Via.

I didn't know that a motherboard maker could actually change how a chipset behaves. And even if you tell me and it's a somewhat valid explination, I still have my doubts. Anyway, I guess that could explain why some KT133A boards are bad and some are much better.

At first, I thought that it was a crappy board with a crappy chipset on it. It seems that I have to give more weight to the first statement...

I personnaly would give VIA another chance. I bash them alot, but I did recommend to people KT266A boards (gigabyte) even when SiS 735 and NForce were around. And these seem to be good boards. I had a VIA KT133 board and it was one most stable computers I had.
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
time said:
I don't think anyone thought it was the best board around. I think inconsistency is the watchword with ECS
Inconsistency is probably the key word. FWIW, I also have a ECS K7S5A and have no problems with it...not the best, but not bad either....of course, YMMV.

KT266, so maybe it wasn't all MSI's fault?
I remember seeing owners of other KT266 based boards also complaining about the same problems plauging the MSI board (i.e. losing USB above 137MHz...)

If I had a dollar for every Abit KT7A owner whining about how bad Via is...Mind you, I'm sure the majority of owners were happy, but there's no denying a vocal percentage had endless grief.
Indeed.

Total BS. Don't include all makers and boards in your claim just because you have a dud.
I agree, my KT133A works fine.

Much has been made of the limited PCI bandwidth available in KT133A solutions. This only affected maximum transfer rates on some RAID setups, and was competely irrelevant to almost everyone. Not to be confused with the SB issue, or course, but as I recall that was partly down to certain manufacturers (Abit) fiddling with settings that were reserved by Via.
In terms of the SB issue, by virtue of the nature of their own device, Creative shares in some of the blame too....although, IIRC, tried to pin it completely on Via.

zx said:
I did not see any chipset related problems with other socket A chipsets like the nForceX
Speaking of which

CK
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
zx said:
I didn't know that a motherboard maker could actually change how a chipset behaves. And even if you tell me and it's a somewhat valid explination, I still have my doubts. Anyway, I guess that could explain why some KT133A boards are bad and some are much better.
Go have a look through some of the documentation on George Breeze's site....he talks about a couple of instances were manufacturers programmed registers Via's datasheets said not to touch.

CK
 

Mercutio

Fatwah on Western Digital
Joined
Jan 17, 2002
Messages
22,275
Location
I am omnipresent
I buy only Via motherboards, in an almost unbroken streak since the last BX board I purchased in, oh, early 2000. I've picked up a couple of AMD 751s (would've bought more but the price was never right) and a pair of SiS735s, but it's Via for me other than that.

Historically, I've found SiS to be a HIGHLY inconsistent manufacturer. Maybe they've turned the corner to genuine respectability. I don't know. Don't care much, either. ALi is even a step down from SiS. I don't think I'd buy either one under any normal circumstances.

Anyway, I've probably worked with every non-P4 Via chipset, and I haven't really had problems with any of them.

Even much maligned choices like the KT133 and KT266 have been more than serviceable for me. Machines were built. Machines are stable, even today. Heck, the Shuttle AK32 I've used in budget builds recently is a KT266 product. It just works, performance is good, and I don't seem to get crashy "oops"-es as I have at points in the past.
 

Handruin

Administrator
Joined
Jan 13, 2002
Messages
13,927
Location
USA
I didn't realize you had posted this link. I posted the same thing as news on the front page... It was by chance that we posted the same thing, I wasn't stealing your thunder. :)
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
BTW, zx, there's a PCR file for the ECS available somewhere on this site if your interested in changing some of its registers, just as you can do (via WPCRxxxx) with Via chipsets etc..

Tannin said:
The retail boxed one was OK, we had a small number of those too, and no problem to speak of, but the allegedly "identical" OEM version was appallingly bad.
I'm curious about the packaging of the OEM boards - how does it differ from the retail? Do you mean you buy preconfigured systems (say from like a DELL OEM) and the MSI board was inside....or do you order in "OEM boards" which you would place into a system package your store offers ...or are these boards that any end user who walks into the store can purchase off the shelf?

CK
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
when I had started my reply your post wasn't...but I took too long to post ....curses! I'll never break the big story :cry:

CK
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
sorry, what I meant was:....your post wasn't up Handruin

....I really should be more careful.

CK
 

blakerwry

Storage? I am Storage!
Joined
Oct 12, 2002
Messages
4,203
Location
Kansas City, USA
Website
justblake.com
CityK said:
BTW, zx, there's a PCR file for the ECS available somewhere on this site if your interested in changing some of its registers, just as you can do (via WPCRxxxx) with Via chipsets etc..

Tannin said:
The retail boxed one was OK, we had a small number of those too, and no problem to speak of, but the allegedly "identical" OEM version was appallingly bad.
I'm curious about the packaging of the OEM boards - how does it differ from the retail? Do you mean you buy preconfigured systems (say from like a DELL OEM) and the MSI board was inside....or do you order in "OEM boards" which you would place into a system package your store offers ...or are these boards that any end user who walks into the store can purchase off the shelf?

CK

wpcredit and wpcrset are so cool...

I'm not positive what it stands for, but I would assume "Windows PC register editor" and "Windows PC register setter"


It's nice to feel you have complete control over your motherboard.. or even a PCI card... dispite what the manufacturer wants... i have used H. Oda's tools for the past couple years and I think they're great.
 

Tannin

Storage? I am Storage!
Joined
Jan 15, 2002
Messages
4,448
Location
Huon Valley, Tasmania
Website
www.redhill.net.au
zx said:
Why is :

SiS735 + ECS (cheap crappy shit) = good board

KT266 + MSI (mi-range?) = crappy board

KT133A + Abit(variable quality?) = crappy board

Tannin, have you tried other Socket A chipsets?

Also, if you can't notice that the KT133A's ATA controller is bad when you do not burn CD's or copy CD's. With hard drives, it's only slower than other ATA controllers. With CD's the system is unresponsive.

Whoever said the pox-ridden ECS SiS thing was a good board? It wasn't me, that is for 110% certain! OK, I tried just a single pair of them, but both were horribly unstable and unsalable, and in neither case was the manufacturer prepared to provide warranty service. I got rid of the second one after having it in stock, unsalable, for nearly a year just a week or so ago: by the simple method of selling it to an optimist for 20% of what I'd paid for it, on the strict understanding that I would provide no warranty or service whatsoever. I have written at length on these theiving scoundrels before.

MSI KT-266. I think we sold a few, maybe 6 or 8. No problems that I remember. But MSI boards are always a lottery. Some models are great, some models are pox. I guess they do maybe 4 models out of five that go OK, then a dud, then three more good ones, another dud, and so on. With MSI, unless you know the exact model already, you pays your money and you takes your chance. Woops! That's the KT-133A I'm thinking of. Never sold an MSI KT-266, can't comment, except to say "see the MSI good:pox ratio above".

Abit. I had a box of 10 Abit KT-133s once. They worked just fine. No failures, all still in service, so far as I remember. Don't think I ever tried an Abit KT-133A. Not to rememmber, anyway. But from what the guys tell me, they are like MSI, only the good board.pox ratio is more like 50/50.

Socket A chipsets I have tried. Well, all of them, I think. Or pretty close.

KT-133: Excellent results with Soltek, FIC and Epox in large quantities (>100 of each). No particular problems that I remember with the odds and ends of other brands we used: ASUS, Abit, A-Open, some others too I think.

KT-133A: Superb chipset: fast, fuss-free, easy as faling off a log. We sold masses of Solteks, quite a lot of Epox ones too, plus various other things No problems to report at all, bar those dreadful MSI ones I already mentioned.

KT-266: Don't think I've ever used one. They weren't around for very long, and there was no point in buying a KT-266 when you could get the same performance out of a KT-133A at half the price, or spend just a little more and get an AMD-760.

KT-266A: Another no-brain best-of-breed chipset, like the 133A. Sold them in similar quantities, still sellling them now. Mostly Soltek & Epox (100s of each), MSI and Gigabyte (maybe 20-odd of each), probably some other ones I forget. Utterly fuss-free, except for some RAM compatibility from the MSIs - but that's what you expect with MSI - and two or three mysterious in-service failures with the 20 Jetways we sold (which, seeing as Jetway is another name for ECS, serves me right).

KT-333: Ho-hum. Sold 30 or 40, mostly Epox and Soltek. Struggled to work out why we were paying the extra as compared to a 266A actually.

KT-400: More ho-hum, unless you are running a 166MHz front-side Athlon, which nobody is yet. Our standard higher-end chipset right now: cheap, simple, practical, no problems that I know about. Not sold that many: 30-odd, nearly all Epox, but a few others in ones and twos.

AMD 750: Sold quite a lot of Gigabyte ones, maybe 200, plus (I think) some others as well. (FIC? Was the AZ-11 one of FIC's hybrid AMD/VIA boards? Or am I thinking of the SD-11?) Apart from having to remember to load different drivers (because it wasn't a VIA chipset but an AMD one) these worked well for us. We put a lot of Durons into Gigabyth AMD 750s.

AMD 760: A small number back when they were the best thing since sliced bread and DDR was like gold bricks only dearer: MSI, Gigabyte, possibly ASUS. A hell of a lot since Gigabyte started specialing them out at a price that just can't be beat, over 100 easily now. Just as stable as a KT-266a or KT-133A, maybe even better, but can sometimes have compatibility problems with particular games or video cards. Not very often though. May be a particular Gigabyte model problem rather than a chipset one, as we didn't get it back in the old days, but then we didn't have today's monster games and video cards either. In either case, pretty rare. Nice chipset.

ALI IforgetwhattheycalleditbuteveryonesaiditwasprettyordinaryandIagree: Sold a tiny number of ASUS boards chith this chipset. Regretted it. But it was probably nothing to do with Ali and everything to do with ASUS - seeing as their KT-133 boards were just as bad. Never failed, just took forever to sort out their plug & pray problems.

SiS 735: About 15 or 20 altogether: those two hopeless slabs of ECS pox, and some Leadtek ones. One box of the Leadteks had about 8 duds in it but that seems to have just been a manufacturing glitch as all the others and the (very prompt) repacements have all worked without a hitch.

Nvidia Nforce 1: Didn't try these bar just one to look at until, believe it or not, about 2 weeks ago. Not because I didn't trust them, but because they were rather scarce (not many board makers had one) and didn't fit into our model line-up anywhere. Too dear for a cheap board, too integrated for a quality board. Who'd buy one? We put them on the price list for a while, but after about three months gave up and took them off again with the sales total standing at zero. But now we are using the ASUS one as an entry-level board. Early days for it, but seems OK so far bar the fact that Kristi says the integrated all-in-one driver flat-out fails to work and she has to fiddle about loading everything by hand, one at a time. But that is probably the ASUS touch. I won't blame Nvidia for it.

Nvidia Nforce 2: Just one, or maybe 2. Kristi hasn't complained, so they must be OK. Not much demand at present. They have to get a bit more serious about their pring if they want anyone to buy one. Right now, unless you happen to want the Firewire or the built-in sound, you can have an Nforce 2 with a Duron 1300 or a KT-400 with an XP 2100. Same price. If you already have a good soundcard, that's a no-brainer.

Also, if you can't notice that the KT133A's ATA controller is bad when you do not burn CD's or copy CD's. With hard drives, it's only slower than other ATA controllers. With CD's the system is unresponsive.

You have a bad board, ZX. We have sold hundreds of KT-133As, and probably 50% or more either shipped with a burner or have had one fitted since. We don't do anything fancy, just bolt the drive in, tick DMA, and go on to the next job. They work. Simple as that. Or you can load the 4-in-1s if you prefer. Lately we have started doing it that way. Doesn't make any difference, they still just work. People come back, buy more blank discs, that's it.
 
Top