Windows 2000 sucks! :(

[Gurm]

Greebe:

Here I have an agenda. For you, in particular, on your machine, you can do whatever the hell you want and I wish you the best of luck - and I might even have some tips for you.

But for Joe Average User, I want them to clearly understand that with an SS7 board they're starting out at a disadvantage and going downhill fast.

Every person I can discourage from buying one of those boards is another monetary loss that gives VIA and ALI incentive to get off their asses and MAKE A GOOD CHIPSET.

- Gurm

------------------
Listen up, you primitive screwheads! See this? This is my BOOMSTICK! Etc. etc.

[The Rock]

I gotta side with Gurm here. Whoever said even the greatest hardware needs patches missed the point. We're talking about chipsets here, not video cards. And the BX chipset has never needed any stupid patches or drivers to make it work. BX till I'm dead baby.

The Rock

[Gurm]

Look, let's try to follow this, ok?

There is an AGP SPECIFICATION.

VIA and ALI did NOT FOLLOW THE SPECIFICATION.

Therefore, all modern high-powered AGP cards (i.e. GeForce, G400, Rage MAXX) crash these chipsets. Thus the chipsets need drivers which disable the advanced features, or slow down some wait states, or something.

The BX chipset follows the spec. How hard is this to understand? Yes, it certainly helps that Intel WROTE the spec. But it is a published spec.

It's like in software - when a company fails to conform to the Win32 API, and then a new OS upgrade crashes the program. Well, if they had followed the API SPEC, they wouldn't have that problem. But instead they blame Microsoft, when in reality their program was noncompliant all along.

And Win95, for example, did not support Pentium II chipsets. They needed drivers. But the OS WORKED with generic drivers - something you can't say for VIA or ALI. And the drivers, when they did arrive, worked first time, out of the gate. The only thing subsequent releases did was make them FASTER.

ALI and VIA still have yet to release a good all-in-one driver set, and their chipsets are over 2 years old now. How much longer will we have to wait for their "driver support", as you call it? How many bad revisions of "drivers" does it take to convince you that the hardware was flawed from the beginning?

- Gurm

------------------
Listen up, you primitive screwheads! See this? This is my BOOMSTICK! Etc. etc.

[Greebe]

Gurm, not to set you off anymore, but WTF dude? Did I say I was a proponent of VIA, NO.
I'm just giving my aged and real world knowledge on the subject.

Rock, Intel couldn't even get their chipsets prior to the BX working 100%. Although the problems weren't as pronounced as VIA's, that's for sure. Even now there latest chipset (MB's) has been recalled do to problems. If you like the BX fine... but the BX won't last for much longer (their only saving grace) and don't you think the this thread has drifted WAY OFF BASE?

[glarec]

do you guys know what the most funny thing is? the original poster (avron) never even post after the init post. :P

[The Rock]

Greebe:

I'm not defending Intel, just the BX chipset. It's too bad Intel wouldn't just redo the BX chipset into a BX+ or something. Intel's latest efforts have sucked monkeys, and VIA appears to finally be on the right track.

P.S. Was this thread ever ON base in the first place?

The Rock
www.3dforce.com

[Greebe]

I have no idear?! LOL

[Wombat]

Rock. Yes, actually, the BX did need patches. But it's been out so long that the lastest versions of Windows have had those patches included by default. It's not necessarily a better chipset -- anything out that long is bound to be well supported. Add to that that MS and Intel collaborate, and try to make each other look good.
As far as publishing standards goes: Intel is no angel. They don't like giving up that information, and I really doubt that their own chipsets would be considered compliant with the AGP specifications. A case in point would be when the i740 cards came out: VIA needed to revise their chipset because the i740 didn't follow the AGP spec (but it worked fine with Intel MB's ... hmmm...).
MS is even worse. They invented SMB, publish the standard, and don't even come close to following their own specifications.

glarec: Drivers or BIOS? I haven't had an Intel MB in my personal computer since 430TX. I had an HX board, and it was a real bitch. Lots of BIOS flashing there. The TX MB was a late manufacture (factory BIOS was stable), but it took a lot of futzing around with Windows to get the proper drivers working.

IMHO, most people who think Intel's chipsets don't have problems can't remember actually dealing with Intel chips when they were new. Just look at the i820. People seem to be having some troubles already. Don't forget that Intel had to cripple the chipset just to get it to market -- it's supposed to handle more RIMM's that it does, but too many glitches, so MB's can't have more than 2(?) sockets on there.


[This message has been edited by Wombat (edited 27 April 2000).]

[Targon]

Ok, I'm also going to jump in on this one, since I've run MOST of my socket-7 systems with VIA chipsets, going from the VIA VP2, to the VP3, to the MVP3. Yes, there have been AGP issues from time to time. BUT, the most obvious case of a problem being caused by Microsoft is the case of the VP3 and MVP3 chipsets. Many are not aware that with these chipsets, Win95 and Win98 would improperly assign IRQs to the cards in the system, and IRQ shareing, which WORKS well in the BIOS would encounter all sorts of problems under 95/98. The reason for this is that Win95/98 do NOT talk to these chipsets properly. If you then install the IRQ miniport driver that VIA released, then the system becomes as stable as a rock, no crashes, or incompatability problems at all. This is in no way a problem with the chipset, but is a problem with Microsoft trying to be too intelligent, and not working properly.

Times have changed a bit, socket-7 is old, and the Athlon is the CPU to go for if you don't want Intel(or if you prefer better floating point performance). The AMD 750 chipset, while lacking features, is very stable for many people. In fact, most problems that Athlon users encounter are caused by power supplies, and memory that are far from being quality products. The FIC SD11 mobo is also the biggest piece of junk I've ever owned, and I replaced it within a few weeks because of the problems. But, quality control issues of MOTHERBOARDS aside, the chipset itself is fine, and does the job. It's not GREAT, but AMD never expected to go into the chipset business for more than a few months. As for the new KX133 chipset, we will need to wait and see if there are any problems that are really chipset related.

So, how does this relate to Matrox video cards? It's simple, if your OS doesn't know how to talk to the chipset PROPERLY, it may do things like mis-allocate IRQs. In the case of the MVP3 chipset, if you don't install the drivers, then the OS won't talk to the vid card properly, and things will break. It's NOT an AGP issue, it's an OS issue.

[Rags]

Most of the problems relating to SS7 systems has been the noisy bus issue. Most BX boards do not have a problem with bus noise, but if die hard SS7 users want to sit here and tell us that the problems are due to Matrox, Intel, MS, etc. etc. that's fine, but Matrox, Intel, etal. are not the ones who made a noisy motherboard, and the only ways to make up for the motherboard noise problems is to retard the timings of the memory, AGP transfers, DMA transfers, etc. I find it funny that people come here (like Hawkeye) and say "Well, the BX did need patches at times, the only reason it is doing well is because it was out longer...". Well, the BX chipset was out about the same time as SS7 came about, maybe a bit earlier, and the BX chipset has had ONE patch needed, not "Oh, I hope intel releases a better driver this week..." like SS7 users have had to go through, I know, because I am a former SS7 user myself, and I just recently offloaded that sucker onto someone else, and I have never looked back. Hopefully the chipset mfrs. and mobo mfrs. get their sheet together and start producing fast, stable, and compatible solutions, or we PC users are going to suffer tremendously (this goes for intel as well, their new chipsets wreek of vx and hx all over again). Hopefully they can learn from the BX chipset and motherboards and implement that into their upcoming designs to wow us again like they had before.

Rags

[Phips]

Are non-Intel chipsets flawed?

Let's try the elimination approch.
If they are not working properly it can be one of two ways:
1- Bad timing and/or noise on the IOs.
2- Register level (ie different configuration that your OS/Drivers don't deal with the right way).

If no.1 is true, then it's like lottery: the non-Intel MB you buy can be buggy or not, it would be a question of luck because timing parameters vary from chip to chip. But, since most of the "pro" (Athlon, SS7, etc...) users on this site allways get their system working after a while, I don't think that's the cause.

If no.2 is true, then it's clearly a software fault. It would mean MS, Matrox and others are not spending as much time testing on those systems as they are on Intel chipsets.
If that is the case, it will be corrected later in time because there are more and more non-intel MB out there.

It would be great to have a database of what to install with whatever OS, MB and Matrox card in here...

[Rags]

We do have a knowledge base, it's called the FAQ forum and the search function

[LAMFDTK]

You seem like a pretty smart guy....

[Gurm]

Phips:

What you're overlooking is that most of the "pro users" eventually get their SS7 chipset to install by turning off features. Either they disable AGP2x, or they add more wait states, or they turn off caching of some sort.

This indicates that it is not only possible but likely that there is a problem with the boards themselves, not the drivers or the testing.

- Gurm

------------------
Listen up, you primitive screwheads! See this? This is my BOOMSTICK! Etc. etc.

[everettes]

Let me chime in. The ONLY problem I've ever had with ANYTHING Intel was the first month I had the CC820. NOW, we're talking about a NEW board (CC820), NEW O\S (Windows 2000), and mostly NEW (GREEN) drivers from ALL manufacturers. A lot of people put the CC820 down. I've found with ALL Intel boards (and I KNOW Gurm will verify) that if you IGNORE Intels specs for RAM, you're asking for it. Once I removed my SE440BX-2 tested SDRAM (Kingston ECC @ $300 per 128 MB Stick that I ASSUMED would work in the CC820)and got some Crucial CC820 Full Tested SDRAM 80% of my problems disappeared. Once Adaptec and Matrox came out with Certified drivers the rest of my problems disappeared. No patches, no cheap hardware, no problems. I'm happy to say that this is the nicest board I've ever had. The MTH may cause a performance hit, but at 600 Mhz It's invisible to me. I built about 10 VIA chipset PC's for a customer that needed them for teaching Preschool (She was worried about price). I lost no less than $1000 on that deal with the plethra of service calls and headaches. Once you've had an Intel chipset (in this age of multimedia performance that is) you'll never go back. And I could care less about an Athlon until they're running on Intel Chipsets. ALI and Via are four letter words when speaking to most support Techs and builders.

------------------
PIII600EB, CC820, 256Megs, 2940U2W, Seagate Cheetah,G400-32 SH, Plextor 8\20, Plextor 40Max, SBLive!, MS FFPro, Altec Lansing ACS-48, Sony CDP 520GS