The following link may explain why people used to encounter one of the many issues people used to encounter back in the day when running VIA chipsets:
VIA Chipsets slow down PCI cards
If you look at some of the pro sound card or video capture cards, at least the ones from back in that era...and maybe even now who knows, you'll see that for some manufacturers for some cards, they actually recommend or have warnings about VIA chipsets. I really wonder if it isn't because of what was found in the link above...
For the record, VIA may be super perfect now...but for anyone to say that they were perfectly perfect back in the K6-2/3 era, or the early K7 era, I think would be mistaken. Maybe with high quality parts and the right mix of parts...
...but not with low quality and the full range of products out there...not like the Intel LX and BX based motherboard were.
Flame suit on...
Chuck
VIA Chipsets slow down PCI cards
If you look at some of the pro sound card or video capture cards, at least the ones from back in that era...and maybe even now who knows, you'll see that for some manufacturers for some cards, they actually recommend or have warnings about VIA chipsets. I really wonder if it isn't because of what was found in the link above...
For the record, VIA may be super perfect now...but for anyone to say that they were perfectly perfect back in the K6-2/3 era, or the early K7 era, I think would be mistaken. Maybe with high quality parts and the right mix of parts...
...but not with low quality and the full range of products out there...not like the Intel LX and BX based motherboard were.
Flame suit on...
Chuck
Zap
Elite Member
- Oct 13, 1999
- 22,377
- 7
- 81
The LX and BX chipsets were nice, but it was totally possible BITD to find crappy boards with those chipsets. One example off the top of my head were these LX chipset pseudo-dual slot 1 boards by Amptron. Bleh! Besides many outright DOA and eventual failures, they were really picky with the CPU having to be inserted just right or no POST.
chucky, "VIA chipset slows down PCI cards"? No. Crap programming of VIA chipsets makes their PCI bus slow. This is because VIA's PCI engine used to be highly programmable - and if the BIOS writers for a particular mainboard didn't care, then either performance went or reliability.
It is this programmability that ultimately let the SB!Live work despite its (!) PCI bug.
That tecchannel article neatly demonstrates that point (if you know enough about PCI to see it): The originally observed poor performance is because the tested setup has the balance shifted to extreme fairness (no single card gets to own the bus for long - page 8). This was good for the early years of PCI, where cards didn't have much throughput but were kind of twitchy on latency because of their I/O buffers being small.
Shift the chipset programming more toward single-card performance, et voilà, great PCI performance. Pages 11 through 13 of the review. It's been four years, BIOS programming has improved since. With bus saturating cards like UDMA/133 IDE, PCI setup is always a compromise - give the IDE card throughput and your TV or sound card might stutter. Give the media cards a regular timeslice, and your IDE throughput suffers. No matter what chipset - but at least with VIA's you get a choice.
Zap, that Amptron (PC-Chips) pseudo-dual-CPU board actually did do dual CPUs if you managed to get hold of the little "APIC" module. The Intel chipsets of the day needed that little helper chip, it wasn't integrated into the southbridge yet.
It is this programmability that ultimately let the SB!Live work despite its (!) PCI bug.
That tecchannel article neatly demonstrates that point (if you know enough about PCI to see it): The originally observed poor performance is because the tested setup has the balance shifted to extreme fairness (no single card gets to own the bus for long - page 8). This was good for the early years of PCI, where cards didn't have much throughput but were kind of twitchy on latency because of their I/O buffers being small.
Shift the chipset programming more toward single-card performance, et voilà, great PCI performance. Pages 11 through 13 of the review. It's been four years, BIOS programming has improved since. With bus saturating cards like UDMA/133 IDE, PCI setup is always a compromise - give the IDE card throughput and your TV or sound card might stutter. Give the media cards a regular timeslice, and your IDE throughput suffers. No matter what chipset - but at least with VIA's you get a choice.
Zap, that Amptron (PC-Chips) pseudo-dual-CPU board actually did do dual CPUs if you managed to get hold of the little "APIC" module. The Intel chipsets of the day needed that little helper chip, it wasn't integrated into the southbridge yet.
Zap- I'm not arguing there weren't/aren't cr@p Intel motherboards out there...we all know we can find some value manufacturers that sell the absolute cheapest motherboards on the planet, and chances are they aren't going to be as compatible, stable, and/or upgradeable as others out there - it comes with the territory of buying price over perfomance or stablility, or a mixture of the two.
All I'm saying is that, in my opinion, the average VIA board of the day back in the MVP3 and early K7 days was not at the level of overall compatibility or stability (not my PC is on for 2 hours and then I turn it off...real stability) that the average Intel boards were.
I'm guessing/hoping that's changed now, and that VIA finally got their act together....but, after going through the often nightmare and/or frustration that was MVP3, and seeing what the early Athlon adopters went through - and I'm not knocking the early Athlon's - again, in my opinion, this is where VIA got their reputation from.
Peter- I just named the link exactly what the authors named their article for consistency's sake.
It's all fine and well if VIA's PCI BIOS programming is highly configurable, but if VIA doesn't/didn't provide the necessary support to the engineers and/or BIOS programmers for the companies using their chipsets, and those motherboards turned out like they ended up turning out...then really, what's the difference if the chipset itself was bad?
As an end user, it really doesn't matter to me if the average VIA motherboard has a great chipset but bad BIOS programming, or, if the chipset blows but the BIOS programming is great, or, if the chipset sucks and the BIOS programming sucks: The end result is the same, I experience instability and/or incompatibility.
If Intel puts out a chipset that has less flexibility that results in greater stability, I have a great deal of confidence that almost all users would rather want that over great BIOS programming flexibility (that they'll never use, since their end users) but the greater chance of instability/incompatibility.
The point of that article, and the heart of the SB Live! and real pro media (sound, video capture) card issues on VIA chipset based motherboards, is that the Intel based motherboards did not exhibit the weight and scope of the problems VIA based motherboars exhibited - end of story.
That an online publication had to put out a multipart article basically showing how bad VIA was compared to the competition, for VIA to even take steps to fix the problem - and, they didn't really fix it, since the patch was only for Promise cards I believe (how does that help the onboard VIA IDE controller?) - just shows how much VIA cared.
The real questions to ask are: Why did the issue ever leave the labs of VIA in the first place? Why when the numerous compatibility and stability complaints started coming in from the user community did VIA itself not solve their problems? Why were their drivers so wacky? Etc.?
I freely admit I have no idea how VIA is now, but anyone who's going to go out and say MVP3 was great really needs to sit down and remember back how it was...
...and for any flamers, check out what I'm running - that is the system I'm writing this on...
Chuck
All I'm saying is that, in my opinion, the average VIA board of the day back in the MVP3 and early K7 days was not at the level of overall compatibility or stability (not my PC is on for 2 hours and then I turn it off...real stability) that the average Intel boards were.
I'm guessing/hoping that's changed now, and that VIA finally got their act together....but, after going through the often nightmare and/or frustration that was MVP3, and seeing what the early Athlon adopters went through - and I'm not knocking the early Athlon's - again, in my opinion, this is where VIA got their reputation from.
Peter- I just named the link exactly what the authors named their article for consistency's sake.
It's all fine and well if VIA's PCI BIOS programming is highly configurable, but if VIA doesn't/didn't provide the necessary support to the engineers and/or BIOS programmers for the companies using their chipsets, and those motherboards turned out like they ended up turning out...then really, what's the difference if the chipset itself was bad?
As an end user, it really doesn't matter to me if the average VIA motherboard has a great chipset but bad BIOS programming, or, if the chipset blows but the BIOS programming is great, or, if the chipset sucks and the BIOS programming sucks: The end result is the same, I experience instability and/or incompatibility.
If Intel puts out a chipset that has less flexibility that results in greater stability, I have a great deal of confidence that almost all users would rather want that over great BIOS programming flexibility (that they'll never use, since their end users) but the greater chance of instability/incompatibility.
The point of that article, and the heart of the SB Live! and real pro media (sound, video capture) card issues on VIA chipset based motherboards, is that the Intel based motherboards did not exhibit the weight and scope of the problems VIA based motherboars exhibited - end of story.
That an online publication had to put out a multipart article basically showing how bad VIA was compared to the competition, for VIA to even take steps to fix the problem - and, they didn't really fix it, since the patch was only for Promise cards I believe (how does that help the onboard VIA IDE controller?) - just shows how much VIA cared.
The real questions to ask are: Why did the issue ever leave the labs of VIA in the first place? Why when the numerous compatibility and stability complaints started coming in from the user community did VIA itself not solve their problems? Why were their drivers so wacky? Etc.?
I freely admit I have no idea how VIA is now, but anyone who's going to go out and say MVP3 was great really needs to sit down and remember back how it was...
...and for any flamers, check out what I'm running - that is the system I'm writing this on...
Chuck
Intel vs. VIA vs. Creative: Creative, back then, blatantly stated, "we test with Intel chipsets, and we don't care about everything else".
(And if you believe Intel's chipsets are bugfree or can't be fscked up by poor BIOS jobs, you're hallucinating.)
Crap BIOS programming is something that still happens, and is completely beyond the control of the chipset makers. See also: ASRock's Dual939 and its BIOS mess. Chipset companies do put out BIOS writer's guides, but they don't force BIOS writers to actually read them.
VIA's patch was a quick fix; improved BIOSes appeared all over the place a mere weeks later. A genuinely bad chipset would have had that impossible.
And yes, MVP3 was a real good chipset. It was the first chipset that got VIA into the "performance" arena, and perform it did (if you let it). Isn't it funny that your 1998 mainboard sustained you all the way into 2006?
I don't know how far back and how deep down your insight goes - I for one have been a BIOS writer for ten years straight, and I damn sure know what I'm talking about.
All you've done above is collected all the insightless spin the "informed internet press" had back then. If you can't see the difference between a non-working chipset and less than optimal programming, I really can't help you. Until you present facts that suggest otherwise, I'll file you under Myth Spinners. Or maybe Parrot.
(And if you believe Intel's chipsets are bugfree or can't be fscked up by poor BIOS jobs, you're hallucinating.)
Crap BIOS programming is something that still happens, and is completely beyond the control of the chipset makers. See also: ASRock's Dual939 and its BIOS mess. Chipset companies do put out BIOS writer's guides, but they don't force BIOS writers to actually read them.
VIA's patch was a quick fix; improved BIOSes appeared all over the place a mere weeks later. A genuinely bad chipset would have had that impossible.
And yes, MVP3 was a real good chipset. It was the first chipset that got VIA into the "performance" arena, and perform it did (if you let it). Isn't it funny that your 1998 mainboard sustained you all the way into 2006?
I don't know how far back and how deep down your insight goes - I for one have been a BIOS writer for ten years straight, and I damn sure know what I'm talking about.
All you've done above is collected all the insightless spin the "informed internet press" had back then. If you can't see the difference between a non-working chipset and less than optimal programming, I really can't help you. Until you present facts that suggest otherwise, I'll file you under Myth Spinners. Or maybe Parrot.
Chucky2, Since Intel has designed most of the standards for modern PCs it is reasonable assumption that Intel chipsets are more compatable than most. Even if an Intel chipset and a third party device had a conflict who would blame Intel?
Personally I'm not a big fan of VIA, but I did have a MVP3 based board and even though my sound blaster live did crackle and pop durring boot up once the system was running I had no problem out of it. The biggest problem (IMO) that I had with the MVP3 was with USB support (this was probably more of a windows 98SE problem though). For the super 7 platform I actually thought the ALi chipset provided a smoother and more responsive feeling than the VIA although it had a few limitations of its own if I remember correctly.
I guess I can see why some people dislike VIA, they may have rushed a few products to market and had some quirks due to their design, the thing is I don't think any chipset can claim otherwise.
What I have a hard time with is the blind loyalty some people display twards chipset brands. There are some chipset brands I would prefer but in the end I could be satisfied with any brand out there right now. The thing is my prefered brands don't conform to the majority opinion on most "enthusiast" sites.
Personally I'm not a big fan of VIA, but I did have a MVP3 based board and even though my sound blaster live did crackle and pop durring boot up once the system was running I had no problem out of it. The biggest problem (IMO) that I had with the MVP3 was with USB support (this was probably more of a windows 98SE problem though). For the super 7 platform I actually thought the ALi chipset provided a smoother and more responsive feeling than the VIA although it had a few limitations of its own if I remember correctly.
I guess I can see why some people dislike VIA, they may have rushed a few products to market and had some quirks due to their design, the thing is I don't think any chipset can claim otherwise.
What I have a hard time with is the blind loyalty some people display twards chipset brands. There are some chipset brands I would prefer but in the end I could be satisfied with any brand out there right now. The thing is my prefered brands don't conform to the majority opinion on most "enthusiast" sites.
Peter- I'm under no such delusions about the greatness of Creative's products...or their lovely drivers they put out...I'm not making exuses on behalf of VIA for Creative's lack of testing for the SB Live! product....for sure Creative shared blame.
I'm also under no delusions that Intel's products are bug free...anyone care to remember i820?
The fact though is that many VIA based users had problems with MVP3 (disproportionately more than Intel based users of chipsets of that era), and pro audio and video capture folks also had disproportionately more problems with their add-in cards on VIA chipset based systems than Intel chipset based systems of that era.
Without going back and re-familiarizing myself with the exact timeframe on the "fix" by VIA, I'll off the cuff say that VIA didn't in any way quickly come out with a fix. Again, it took people posting about George Breese's PCI latency patch all over the net - including on what was VIAHardware.com (which is now I think Sudhain or something like that) and VIA's own VIAArena.com forum's - and the article I linked to above (which is finally what shamed them IMHO) for them to finally issue a patch. A patch for something they themselves should have been making sure performed correctly on motherboards going out to end users*.
Keep in mind that the patch VIA issued doesn't help anything but what, Promise IDE add-in cards? What about the pro audio cards, and the video capture cards, or the other PCI add-in cards that had problems (especially the ones where George's patch helped with)? Guess it's a good thing that George Breese was around to offer his "0 PCI latency" patch, or whatever the heck it was called. A patch ironically that solved many a peoples problems (or at least significantly reduced them) and that VIA warned against on their own VIAArena forum, as people were/are begging them to help them get their hardware working on VIA based products.
MVP3 is a good chipset, although the memory performance in comparison to the ALi Alladin V was not as good...nor the IDE performance if I remember correctly. And Yes, we can all probably agree that the USB performance wasn't super stellar either. As I said and you pointed out, I'm still on a MVP3 based system today...so good MVP3 based systems do exist and are common. But that's not what this thread topic is about...
I'm very happy that my late 98/99 mainboard has sustained me into 2006...the Asus P55T2P4 using an Intel HX chipset and a K6-3 450 (6x75) at my relatives house has for them as well...should I applaud Tyan and VIA for making the AGP, PCI, and IDE on this Tyan S1598C2 work as specified in those standards? Isn't that what their supposed to do? I'm pretty sure I shouldn't be grateful for that...that's the minimum that we should all expect from products that we buy..especially considering this mainboard was $140 USD or somewhere in that area when I bought it back in 2000...that wasn't exactly budget back then.
My insight goes back to about VIA VP3 days...and then definitely a good amount of experience with MVP3. I completely respect your Elite status, and your experience as a BIOS writer for 10 years, but I'm not sure how that's factoring into this discussion? That VIA's BIOS is uber programmable means nothing IMO... I as an end user do not program BIOS's...I take motherboards that are supposed to work, and plug - where appropriate - add-in cards that are supposed to work into these motherboards and then install an OS, drivers, and I should be off and running.
I accept that every piece of hardware out there cannot work with every other piece of hardware...Lord knows I've ran into perfectly working cards that refused to work on Intel based motherboards when they were put in those systems....just, far less likely than I did when doing swaps over to VIA based systems.
You claim I can't see the difference between a non-working chipset and less than optimal programming...and what your not seeing is that I - as an end user - do not care. If I pick up 100 different Intel or nVidia or SiS or ALi based motherboards, and 100 different VIA based motherboards, and using all the same other components, out of those 100 different builds for each, chipset brand X has 10 hassles overall and VIA has 25 hassles overall, it matters none to me whether or not the VIA chipset is inherantly at fault, or the BIOS programming: All I know is that I'm not going to go the VIA route next time, period...end of story.
*I seriously wonder sometimes why VIA would not add in their supply contracts with large vendors that the final product would undergo a round of performance and compatibility testing by VIA to ensure the final product met a minimum set of performance and compatibility guidelines. Even if VIA didn't have the muscle to get this into their contracts, the last they could have done - in the interest of creating and maintaining a good rep in the community - would be to aquire the final product boards as we all do and then conduct their own round of testing. This way, if the ID'd an issue, they could notify the mainboard manufacturer and ideally even provide them with a fix. I canot imagine a mainboard manufacturer getting approached by the chipset vendor that the vendor has a recommnendation for a fix for a defect they think they found, and not try and implement the fix (providing of couse it wasn't a mass recall, in which I have no doubt many/most mainboard manufacturers would just bury the info and issue a "revision"). If your VIA you can easily see which mainboards based on your product are taking off and take steps to insure at least those products perform as you as a company want your rep. to be tied to.
The opposite of this seems to be the industry standard for far too many companies though...which is really a shame, since I really doubt it'd cost these companies that much overall to implement this kind of external internally driven QC.
I have no idea who Parrot is...
justly- I completely understand and agree with all you've posted, except I have had to deal with problems out of the MVP3 based systems - which does not mean they can't be made to be stable...albeit sometimes with some work.
I don't have brand loyalty either...I simply use Intel in my posts above because that was VIA's main competitor at the time I feel they dug their own reputation...ALi had the Alladin V, but while a good chipset, did not really take of like the MVP3 did.
Folks, VIA motherboards may be the best out for end users today for all I know...but the OP started a thread on why people badmouth VIA...and I'm just relaying my position on why VIA has the rep or the stain on their rep that they do. In the end, it's JMHO...
Chuck
I'm also under no delusions that Intel's products are bug free...anyone care to remember i820?
The fact though is that many VIA based users had problems with MVP3 (disproportionately more than Intel based users of chipsets of that era), and pro audio and video capture folks also had disproportionately more problems with their add-in cards on VIA chipset based systems than Intel chipset based systems of that era.
Without going back and re-familiarizing myself with the exact timeframe on the "fix" by VIA, I'll off the cuff say that VIA didn't in any way quickly come out with a fix. Again, it took people posting about George Breese's PCI latency patch all over the net - including on what was VIAHardware.com (which is now I think Sudhain or something like that) and VIA's own VIAArena.com forum's - and the article I linked to above (which is finally what shamed them IMHO) for them to finally issue a patch. A patch for something they themselves should have been making sure performed correctly on motherboards going out to end users*.
Keep in mind that the patch VIA issued doesn't help anything but what, Promise IDE add-in cards? What about the pro audio cards, and the video capture cards, or the other PCI add-in cards that had problems (especially the ones where George's patch helped with)? Guess it's a good thing that George Breese was around to offer his "0 PCI latency" patch, or whatever the heck it was called. A patch ironically that solved many a peoples problems (or at least significantly reduced them) and that VIA warned against on their own VIAArena forum, as people were/are begging them to help them get their hardware working on VIA based products.
MVP3 is a good chipset, although the memory performance in comparison to the ALi Alladin V was not as good...nor the IDE performance if I remember correctly. And Yes, we can all probably agree that the USB performance wasn't super stellar either. As I said and you pointed out, I'm still on a MVP3 based system today...so good MVP3 based systems do exist and are common. But that's not what this thread topic is about...
I'm very happy that my late 98/99 mainboard has sustained me into 2006...the Asus P55T2P4 using an Intel HX chipset and a K6-3 450 (6x75) at my relatives house has for them as well...should I applaud Tyan and VIA for making the AGP, PCI, and IDE on this Tyan S1598C2 work as specified in those standards? Isn't that what their supposed to do? I'm pretty sure I shouldn't be grateful for that...that's the minimum that we should all expect from products that we buy..especially considering this mainboard was $140 USD or somewhere in that area when I bought it back in 2000...that wasn't exactly budget back then.
My insight goes back to about VIA VP3 days...and then definitely a good amount of experience with MVP3. I completely respect your Elite status, and your experience as a BIOS writer for 10 years, but I'm not sure how that's factoring into this discussion? That VIA's BIOS is uber programmable means nothing IMO... I as an end user do not program BIOS's...I take motherboards that are supposed to work, and plug - where appropriate - add-in cards that are supposed to work into these motherboards and then install an OS, drivers, and I should be off and running.
I accept that every piece of hardware out there cannot work with every other piece of hardware...Lord knows I've ran into perfectly working cards that refused to work on Intel based motherboards when they were put in those systems....just, far less likely than I did when doing swaps over to VIA based systems.
You claim I can't see the difference between a non-working chipset and less than optimal programming...and what your not seeing is that I - as an end user - do not care. If I pick up 100 different Intel or nVidia or SiS or ALi based motherboards, and 100 different VIA based motherboards, and using all the same other components, out of those 100 different builds for each, chipset brand X has 10 hassles overall and VIA has 25 hassles overall, it matters none to me whether or not the VIA chipset is inherantly at fault, or the BIOS programming: All I know is that I'm not going to go the VIA route next time, period...end of story.
*I seriously wonder sometimes why VIA would not add in their supply contracts with large vendors that the final product would undergo a round of performance and compatibility testing by VIA to ensure the final product met a minimum set of performance and compatibility guidelines. Even if VIA didn't have the muscle to get this into their contracts, the last they could have done - in the interest of creating and maintaining a good rep in the community - would be to aquire the final product boards as we all do and then conduct their own round of testing. This way, if the ID'd an issue, they could notify the mainboard manufacturer and ideally even provide them with a fix. I canot imagine a mainboard manufacturer getting approached by the chipset vendor that the vendor has a recommnendation for a fix for a defect they think they found, and not try and implement the fix (providing of couse it wasn't a mass recall, in which I have no doubt many/most mainboard manufacturers would just bury the info and issue a "revision"). If your VIA you can easily see which mainboards based on your product are taking off and take steps to insure at least those products perform as you as a company want your rep. to be tied to.
The opposite of this seems to be the industry standard for far too many companies though...which is really a shame, since I really doubt it'd cost these companies that much overall to implement this kind of external internally driven QC.
I have no idea who Parrot is...
justly- I completely understand and agree with all you've posted, except I have had to deal with problems out of the MVP3 based systems - which does not mean they can't be made to be stable...albeit sometimes with some work.
I don't have brand loyalty either...I simply use Intel in my posts above because that was VIA's main competitor at the time I feel they dug their own reputation...ALi had the Alladin V, but while a good chipset, did not really take of like the MVP3 did.
Folks, VIA motherboards may be the best out for end users today for all I know...but the OP started a thread on why people badmouth VIA...and I'm just relaying my position on why VIA has the rep or the stain on their rep that they do. In the end, it's JMHO...
Chuck
I had an ALi based motherboard back in Super7 times (the P5A) and briefly switched to a VIA based board - worst mistake I ever made, the VIA board was trash compared to the ALi board (and it was a much hyped epox model at the time). ALi's biggest problem back then was their AGP driver (which was eventually fixed, but took forever). I quickly went back to the ALi board.
I was also an early adopter of Socket A (having sat out Slot A) and used both KT133 and KT133A (this involved the MSI motherboards that Anandtech pimped and turned out to suffer from crap-ass-itors - neither board still works). I know the TV card problems that were supposedly fixed were never fixed for me no matter what I tried. Needless to say I switched to nForce1 virtually the day it was released and nForce1 absolutely slaughtered the VIA boards IMHO.
You wouldn't convince me to use VIA again, no matter what so long as there is another chipset available for use.
I was also an early adopter of Socket A (having sat out Slot A) and used both KT133 and KT133A (this involved the MSI motherboards that Anandtech pimped and turned out to suffer from crap-ass-itors - neither board still works). I know the TV card problems that were supposedly fixed were never fixed for me no matter what I tried. Needless to say I switched to nForce1 virtually the day it was released and nForce1 absolutely slaughtered the VIA boards IMHO.
You wouldn't convince me to use VIA again, no matter what so long as there is another chipset available for use.
I've always liked VIA myself.
Even had MB's based on the scary KX133 and KT133 chipsets, never myself had an issue with either!
Even had MB's based on the scary KX133 and KT133 chipsets, never myself had an issue with either!
VIA's got a bad rap because they've made some sketchy chipsets. Plus their 4-in-1 drivers were crap for awhile (are they still? I'm not sure... Haven't seen a reason to go back: i865/i875 was better for Intel P4, NF2 was beter for AMD Athlon XP, NF4 and now RD580 are better for AMD A64/X2/Opteron).
While, for example, the ASUS P3V4X was a VIA P3 slot-1 board that was best in its class (before Intel's 815 chipset came out), VIA had several missteps/revisions of chipsets after that.
The KT100 and KT133 were solid if unspectacular chipsets. KT266 was crap and had horrible memory performance. They followed it up with the better KT266A and KT333 chipsets and the once again lackluster KT400 chipset. When Nforce 2 came out, it blew out of the water anything VIA had ever made for AMD.
While VIA makes functional chipsets, they are almost never "best in class" and are often only purchased because they're cheap and because VIA is quick to adapt to new form factors and sockets, while Intel, Nvidia, etc iron out the bugs on their design before releasing their chipsets (generally speaking; though Nvidia takes more chances on odd-numbered chipsets, eg NF1 and NF3).
VIA has (to my knowledge) always had crap IDE drivers for example (Nvidia is guilty of this as well). While by no means terrible, I've always had better success with rival chipsets to VIA ones. Plus for a company that has been in the game so long, they show remarkable ability to not improve that much - they're still middle of the road (at best).
Nvidia hit it off like gangbusters with the decent Nforce1 and then released a best-in-class chipset with NF2 for AMD. The NF4 was also a smash hit, introducing SLI and other great features.
Having just set up an ASUS A8R32-MVP motherboard, I can say that ATI has also quickly released a superior second-generation chipset. While VIA chugs along as they always have; with decent boards with OK features and mediocre drivers...
While, for example, the ASUS P3V4X was a VIA P3 slot-1 board that was best in its class (before Intel's 815 chipset came out), VIA had several missteps/revisions of chipsets after that.
The KT100 and KT133 were solid if unspectacular chipsets. KT266 was crap and had horrible memory performance. They followed it up with the better KT266A and KT333 chipsets and the once again lackluster KT400 chipset. When Nforce 2 came out, it blew out of the water anything VIA had ever made for AMD.
While VIA makes functional chipsets, they are almost never "best in class" and are often only purchased because they're cheap and because VIA is quick to adapt to new form factors and sockets, while Intel, Nvidia, etc iron out the bugs on their design before releasing their chipsets (generally speaking; though Nvidia takes more chances on odd-numbered chipsets, eg NF1 and NF3).
VIA has (to my knowledge) always had crap IDE drivers for example (Nvidia is guilty of this as well). While by no means terrible, I've always had better success with rival chipsets to VIA ones. Plus for a company that has been in the game so long, they show remarkable ability to not improve that much - they're still middle of the road (at best).
Nvidia hit it off like gangbusters with the decent Nforce1 and then released a best-in-class chipset with NF2 for AMD. The NF4 was also a smash hit, introducing SLI and other great features.
Having just set up an ASUS A8R32-MVP motherboard, I can say that ATI has also quickly released a superior second-generation chipset. While VIA chugs along as they always have; with decent boards with OK features and mediocre drivers...
Gstanfor, you know that the ALi Aladdin V north bridge took SEVEN chip revisions before it fully worked? That one was plagued with AGP compatibility issues, non-working cache tag RAM, etc. pp. - all of which were actual silicon bugs, not just poor drivers.
It still finished up being able to run the GeForce256 SDR I had at the time better than the VIA board could though... The only real gotcha was on the P5A-B which I sold a fair few of - the AGP slot didn't supply enough power due to cheapo voltage regulator, quite a common problem back then LX or LZ also suffered from it.
There was lots of discussion about this back in the day peter, and nVIDIA, *were* within the AGP Spec, you can download the spec and check out power requirements for yourself vs GeForce power requirement. Even nowadays you can't trust some motherboard makers to supply *all* the power a graphics slot is specified to supply as some 7600GT owners have discovered (no molex on that card to fix the situation either).
Anyway, I probably wasn't a "typical" Super7 user, I had the GeForce (when everyone else liked 3dfx), 320mb memory, K6-III 450, TV Card, A3D sound card. The ALi handled all of that quite well with good fast performance. The VIA board had hardware compatability fits with some of the hardware and simply felt slower to boot.
Anyway, I probably wasn't a "typical" Super7 user, I had the GeForce (when everyone else liked 3dfx), 320mb memory, K6-III 450, TV Card, A3D sound card. The ALi handled all of that quite well with good fast performance. The VIA board had hardware compatability fits with some of the hardware and simply felt slower to boot.
You noticed what my profession is? I read specs all day - TNT2 Ultra was beyond the AGP 1.0 power allowance. Guess why AGP 2.0 was then rushed out the door.
Btw, it took ALi until 2003 to provide a driver patch to make certain TV cards work in direct-to-overlay push mode with their AGP northbridges. And it took their 1543 north bridge five chip revisions before UDMA33 operation /actually/ worked (in rev. C0), so much for speedier boot. And even that has data destroying DMA bugs with drives larger than 128 gigs.
Point being: It's not like VIA is doing worse than anyone else. Different bugs, yes. More bugs, no.
Btw, it took ALi until 2003 to provide a driver patch to make certain TV cards work in direct-to-overlay push mode with their AGP northbridges. And it took their 1543 north bridge five chip revisions before UDMA33 operation /actually/ worked (in rev. C0), so much for speedier boot. And even that has data destroying DMA bugs with drives larger than 128 gigs.
Point being: It's not like VIA is doing worse than anyone else. Different bugs, yes. More bugs, no.
I nenver said a word about TNT-2 Ultra, peter, I was talking about GeForce 256 SDR. In any case the TNT-2 Ultra was still well within the AGP specification and it sounds to me as though you need to reread that specification. AGP 1.0 is irrelevant to this discussion since ASUS P5A, P5A-B, the epox VIA board and Intel LX boards all feature AGP x2 slots.
As for the TV card it always worked perfectly for me on ALi, nothing but trouble on VIA, IDE was definitely quicker than VIA's too. All the ASUS boards I ever sold/used were rev 1.05 (did see a second hand 1.03 years later). When dealing with ASUS/ASROCK, you know a board will go through multiple early revisions, it's best to aviod the board until it has (My ASUS A7N8x Deluxe is rev 1.05 as is my ASROCK 939 DUAL SATA II).
Further info on the AGP troubles of the time, the AGP spec called for certain types of power regulators to be used for the AGP slot. Those boards that gave trouble invariably had that regulator substituted for an underpowered, cheap linear voltage regulator, which is what caused the problem to arise It's also why (most) BX boards never suffered the problem.
As for the TV card it always worked perfectly for me on ALi, nothing but trouble on VIA, IDE was definitely quicker than VIA's too. All the ASUS boards I ever sold/used were rev 1.05 (did see a second hand 1.03 years later). When dealing with ASUS/ASROCK, you know a board will go through multiple early revisions, it's best to aviod the board until it has (My ASUS A7N8x Deluxe is rev 1.05 as is my ASROCK 939 DUAL SATA II).
Further info on the AGP troubles of the time, the AGP spec called for certain types of power regulators to be used for the AGP slot. Those boards that gave trouble invariably had that regulator substituted for an underpowered, cheap linear voltage regulator, which is what caused the problem to arise It's also why (most) BX boards never suffered the problem.
May I recommend you re-read the specs, to recall that AGP mode does not equal AGP standard revision.
AGP 1.0 had 1x and 2x modes, and quite low a power budget. AGP 2.0 added (optional) 4x mode and almost doubled the power allowance. You could make AGP 2.0 boards w/o 4x mode, just by using stronger voltage regulation - but the vast majority of 1x/2x-only boards went by AGP 1.0 specification.
Zap, you nailed it. Most of the perceived "crap chipset" problems are actually the board maker cutting corners - be it in hardware design or software attention to detail.
There was a very easy fix for the AGP issue in the older super 7 boards...just switch the TNT/GF card to AGP 1x mode instead of 2x....problem solved.
I used and still have a Gigabyte GA-5AA rev 1.1 ALi Aladdin V mobo which hate nvidia cards..but it became a real sweetie as soon as you switched it to 1x mode using the ALI AGP utility (the GA-5AA had no 1x/2x option in its AMI bios)
I have even run GF3's and GF4Ti's on this board(as well as my P5A-B) with out any hassles.
I remember when I replaced my GA-5AA with an EPOX EP-MVP3C2 board...an quickly switched back to it after I discovered VIA's horrid drivers and IDE performance..it just wasn't worth the upgrade to a working AGP 2x and ATA66.
I still have all of my Super 7 boards which feature the ALI ALADDIN V, VIA MVP3 and SIS 5598 chipsets...all of them had "issues" but with the drivers setup correctly(I did eventually get that epox working) they worked just as good as any intel 440BX system at the time, even with Nvidia TNT2/GF256/GF2 cards.
I would have to say however that the VIA was my least favorite due to the driver issues and slow IDE perfomance...that soured my taste for VIA right upto the KT133A that was on my ABit KT7A-RAID, that board was an absolute monster with my AMD 1400C. I remember it being faster than many of my friends P4/i850 combo's at the time(early 2001)..and it cost less than half the price.
will889
Golden Member
- Sep 15, 2003
- 1,463
- 5
- 81
True, but i built with those bland DFI boards in a shop along with Soyo and Biostar long before the Lan Party series because we were resellers and got nice bulk deal on them. They were bland but they were tremedously stable. I agree with an above poster about SIS chipsets - they are very good. Maiboard companies simply didn't "trick out" the bios'es and other ammenities on the boards. Had they done so they would have been much more well known and competetive concerning sales for enthusiasts. There is a reason MS chose SIS for the Xbox360 - stability being the first reason (EG non tricked out super hyped always messed with chipsets). SIS just works. They may not be flashy, but people thet do use them then to be non nonsense users that simply need to get work done and don't need instability. In fact over time SIS has been the most stable AMD chipset to date IMO.
MS didn't just choose SiS. They won the contract out (Bidding war). SiS does a whole lot more in the Semiconductor industry than what meets your eye. They've had their core-logics featured in Media boxes as well in the past. Just like VIA has a very large presence in the 1394 chipset market as opposed to Texas instruments and another strong hold in the Audio DSP market as well.
If you think SiS chipsets are more stable than VIA in the AMD market, then your experience must be very short to speak of.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter