Sis745 amd's best bet for upcoming platform???

[Duvie]

I am wondering since all I have read of issues with past via products and now even the touted kt266a platform if via can be counted on to bring a stable bug free system to the market. I have owned past via products and have not seen or had any of the issues but that may have been more to do with lack of certain other hardware items that exasperated the problems in my systems.

I was looking at getting a kt266a cause I knew it was a little bit better performer then my ECS K7S5A mobo. I also wanted some more overclocking features such as multiplier and voltage adjustments. Don't get me wrong my ECS has been great and I have had no hardware issues to date.

I have as of the last few days put my plans of getting a kt266a mobo on indefinite hold. I have no confidence in the product. I feel sorry for AMD who is trying to get a name in households as a stabile platform but have lowsy manufacturers building chipsets that are going to drag the amd name down.

I hope sis can gain more of a foothold with other motherboard manufacturers and get some other 735 systems out as well. ECS I can admit seems to have a little quality issues such as the high number of boards shipped with failing cmos batteries and motherboards shipped with jumpers set incorrectly. Otherwise I think it is obvious the sis chip and its integrated north and south has performed quite well. I have so far not had to update any motherboard drivers unlike vias 4 in 1 drivers that are in a constant stream.


I have built quite a few via systems this last year for others, but no kt266a's. No problems reported yet.


I have recently noticed the disappearance or real lack of athlon xp systems by oem builders listed in the sales advertisements in the paper. I was wondering is if this is cause most stock has been sold or that oems are having a tough time picking a pltform that will be stable for the chip....


What do you think???

What do you thik of the future sis745 offering??

 

[Idoxash]

i was really hoping that the 735 would be cool ( and as it was a nice chipset it seems as the ppl that made the boards just don't know how to make them ) so i hope sis can make some nice, fast, and stable chipsets.....i will not buy via again and as i said in another post if there is not other chipset out in 4 mths that is just as good as intels then i will leve amd cause there is not need to have a kick but chip in your pc if the pc is not up to speed or is stable enought to do any thing on.........my next pc will be a true "computer" not some lame crap from via.

SIS kick some but

 

[AZGamer]

There are issues with the KT266A, but there are issues with the 735/745 as well.

The OC-ability has more to do with the northbridge then the motherboard maker.

As such, 745 will not OC as well as an Ali Magik C-Stepping or Via KT266A.

 

[Duvie]

No!!!!

Ocing on an ecs can be done. They have a bios to allow fsb boosting but the lack of ddr voltage adjustments often wont let most pc2100 users past 146fsb with any stability....This has nothing to do with the northbridge they just don't offer core voltage adjustments which are also needed to achieve any real gains in ocing...

I was just looking for ocing features cause I already have stability with my ecs at default speeds as well as running my system at 1500mhz (thanks to cpucool) right now. I sure as hell don't wish to give up stability and ide performance for the ability to try to oc.

< As such, 745 will not OC as well as an Ali Magik C-Stepping or Via KT266A. >

Maybe so, but it doesn't help if the system is buggy or held back performance wise such as the via's....

i am interested in th ali as well and I am watching the reviews...Hopefully it is a good performer as well...

< There are issues with the KT266A, but there are issues with the 735/745 as well. >

Most of the issues I have seen to date with the ecs is with a certain mobo manufacturers quality control...With via we are talking about fundamnetal flaws in their chipsets. as far as I know there are no bugs like vias occuring in sis chipset. I am not equating this to rma rates of kt266a motherboard maker companies or other manufacturer specific quirks.

 

[oldfart]

Its been the lack of what I consider a bug free platform that has kept me running an Intel system. It has nothing at all to do with AMD. I don't consider VIA a quality product. The PCI latency issue has caused poor IDE performance, soundcard incompatibilities, data corruption, poor USB performance. They have constant 4 in 1 bug patches to try and patch around the issues. With Intel, the very first chipset inf drivers work with the release of a new chipset. Quite a difference. The SiS platforms may be the ones that turn it around. Its a shame ECS was the only one to make a 735 board. We really need the better board mfgrs to produce a high quality 745 board. Until this happens, I'll stick with Intel.

 

[AZGamer]

<< No!!!!

Ocing on an ecs can be done. They have a bios to allow fsb boosting but the lack of ddr voltage adjustments often wont let most pc2100 users past 146fsb with any stability....This has nothing to do with the northbridge they just don't offer core voltage adjustments which are also needed to achieve any real gains in ocing... >>


The only reason I raised this issue was becuase it is one of the things you were thinking about upgrading due to.



<<
Maybe so, but it doesn't help if the system is buggy or held back performance wise such as the via's.... >>


Actually, with the newest 8233 southbridge, 99% of VIA problems are gone that plagued it in the 686B generation, it's only true, buggy, southbridge.



<< i am interested in th ali as well and I am watching the reviews...Hopefully it is a good performer as well... >>



I have had some experience with these, and they bring in a whole new level of problems. It seems the first group of these to market had major design flaws, and were put on KA266+ PCB.



<< < There are issues with the KT266A, but there are issues with the 735/745 as well. >

Most of the issues I have seen to date with the ecs is with a certain mobo manufacturers quality control...With via we are talking about fundamnetal flaws in their chipsets. as far as I know there are no bugs like vias occuring in sis chipset. I am not equating this to rma rates of kt266a motherboard maker companies or other manufacturer specific quirks. >>



So you're saying having to RMA a product is better then just downloading a patch off of the web?

 

[darth maul]

<< Its a shame ECS was the only one to make a 735 board. >>



You may have missed the Leadtek board based on the 735. In fact Newegg carries it. Ya I know its not a Abit, or Asus, or Epox, but atleast someone else made one. To bad VIA heavy handed some of those companies and told them not to make 735 boards, I wonder if some of those companies regret the deciesion to cave into VIA.

With that said, I still might get a 266a, I want to overclock!!!

 

[oldfart]

Someone around here got the Leadtek and said it was a real dud.

 

[RDP]

Every chipset and every motherboard has problems some of the time, and usually they are configuration errors. It's hard to expect that every product will work with every other product, all the time. System builders test and re-test their configs to be sure they work flawlessly, and then don't vary from them. DIY'ers, in contrast, pick willy-nilly and then expect a flawless install. It just doesn't work that way.

I have a K7S5A, and yes, I do have problems, including the CMOS resetting error and memory errors. I blame ECS. However, I have read bad posts about the Leadtek board, as well. Again, no chipset is perfect.

As a result of my K7S5A problems, I just ordered the Gigabyte Kt266A board, which has gotten great buzz as a super stable board. As much as people complain about Via, they update their drivers very frequently, and have a huge base of installed users for tips and suggestions. Remember that with so many Via users, there are bound to be far more unhappy users, number wise, than unhappy users of other chipsets. That doesn't mean the chipsets are bad. It just means that so many people are using so many different variations, it's inevitable that some of them won't work.

 

[oldfart]

Its sad that flaws like this are accepted as "normal". Every chipset does not and should not have problems. Yes, a good chipset made by a poor manufacturer will not end in a good product. However, a flawed chipset will always have problems no matter who makes it. This doesn't help the people with AMD processors, but take the Intel 815, 845, 850 chipsets. Where are all the bugs and flaws that "every chipset has". Those boards are truly rock solid. Use any soundcard you like, any video card you like. The USB works, IDE performance is great. People talk about old "BX stability". The i8xx chips are every bit as stable.

Here is another one that kills me:


<< As much as people complain about Via, they update their drivers very frequently >>


That is viewed as a good thing? They come out with a new 4 in 1 patch every week to try and fix one bug or another. That is ridiculous. I bought a CUSL2 when it first came out. The i815E chipset was brand new. I installed in Intel inf files ONCE. The very first set worked and was bug free. Any new inf drivers have merely added support for newer chipsets such as the 850 and 845. Why can Intel put out stable drivers on the first shot and VIA puts out countless versions and can never get it right?

I really hope the SiS 745 chipset turns out to be a good one. If better board makers like Asus, Abit, MSI, Gigabyte use it, I think there will finally be a good stable platform that I'll be interested in. VIA has disappointed too many times.

 

[Athlon4all]

I do highly regard the recent SiS chipsets (735, 645, and other variants of these 2) and personally, I do agree that while I have (so far) not encourtered any of the big issues that some people have with VIA chipets, I really must question their ability to (still) produce reliable chipsets. They have gotten much better since the MVP3 and KX133 days but still. Right now, I would not say no to VIA chipsets for me, but I must question their reliability. So, we'll see. SiS definately has much to prove in the OEM market (more than VIA has to prove) and I do think that the time is ripe for them to seriously take some ground from VIA.

I do wish that some overclocking boards from Abit, Shuttle, and EPoX would come out, the problems on K7S5A are simply Mobo issues and not inheriant SiS 735 problems, and I might consider getting a Abit KS7 , or a Shuttle or EPoX SiS 735/745 board for my Winter Upgrade, but for now I still think that I am going to plan on getting a Shuttle AK31A (or AK35 GTR). We'll see.

It really for me depends on how soon KS7 is coming out and how much it will cost. You really, cannot argue that at $85 (from Newegg) Shuttle AK31A packs an incredible feature set into an incredibly attractive price, and from what I've heard, AK35 GTR (priced at $85-100) packs even more (Sound, more stability, more performance, and ATA/133 RAID), and really, it will be tough for me to say no to a AK31 or AK35 GTR. To sum up, if performance is close enough, and features/price are similar, I might take a SiS chipset board over a VIA. In other words, I will not say no to VIA because of thier problems. This is certainly something to keep an eye on, but it will be tough but I think that SiS has a very bright future.

 

[oldfart]

<< more performance, and ATA/133 RAID >>


Problem is, because of VIA PCI bus problems the ATA133 RAID will perform substantially worse than any other chipset.

 

[Duvie]

oldfart I believ it was insane3d and he encountered mostly ocing problems and a rather picky board for ram...I think he said it was stable at default just wasn't the great ocing board for sis 735 chipset that it was sold as...

 

[Wildmana]

The only bug that i have read about SIS735 so far is the inability to run FSB and Memory at different speeds i.e. 133/166. For me that is not a problem. The other problems that i read about are the mobo. Two different sites have made the comment that Abit KS7 is going to be a "eye opener". If it is i will have one. I two am done with VIA buggy crap. Although i dont use anything but AMD, they dont have anyone to blame but themself. If they wont make motherboard chipsets then they dont need to cry about the other sorry crap or deserve to be head of the class............

 

[AZGamer]

The reason it's hard to take accusations such as this seriously is they're all anectodal. The fact that VIA has sold more of it's chipsets then other manufacturers means that there will be more complaints in generally, but if you take a close look statistically, the percentage for VIA is probably lower of chipsets that honestly have problems.

If you think that SiS don't suffer from as many problems (if not more), check the OC Workbench Forums. The following are some general issues with SIS 735:

-Inability to run 1,400 MHZ T-Bird at it's 266 FSB
-Inability to correctly interface with memory, leaving many problems with stability
-Sound corruption issues with certain sound cards
-Very few driver updates, if you have a problem your stranded

If you think that Nvidia Chipsets don't have these problems, check the MSI Forums. The following are some general issues with the nForce:

-Sound Capabilites misrepresented
-Onboard sound doesn't allow for movie watching without timing issues
-General inability to support any non-XP OS
-Superstability mode, generally impossible to determine, lowering speed
-Lack of DMA bus-mastering drivers
-Very few driver updates, if you have a problem your stranded

As for VIA? There problems pale in comparison, and there is a wide user community and patches to use if you experience any of the problems.

 

[Duvie]

< -Inability to run 1,400 MHZ T-Bird at it's 266 FSB >

Look at my rig!!! I am running a 1.4 with 266fsb...so I think you are overstating that. I know the issue you are talking about and where you are wrong is that this is an ecs mobo problem not a sis chipset.

You frekkin ppl need to get right what is the chipset and what are things that the individual motherboard companies build...Other sis735 boards like amptron, pcchips, chaintech, & leadtek don't have this problem.

< -Inability to correctly interface with memory, leaving many problems with stability >

Where are you getting this and are you sure once again this isn't a mobo manfacturer issue???

< -Sound corruption issues with certain sound cards >

Which ones??? I have to date used sb pc512, sb live, sb audigy, & snata cruz...some have been on systems I have built for others...I haven't seen any issues to date.

< -Very few driver updates, if you have a problem your stranded >

That is because very little of known problems are actually related to the sis chipset. ECS has released multiple bios upgrades to deal with many issues I think you are confusing with a chipset issue...



Via's bugs are 100% related to the chipset. They can be duplicated on multiple manfacturer's boards of same chipset. They have been duplicated on both intel and amd platforms. That sort of takes the individual manufacturers out of the equation...

 

[AZGamer]

Amptron and PC Chips are both the equivalent of ECS. Neither of their boards are drastically different, and if I remember correctly, they all fall under the same ownership.

Leadtek has a whole new host of issues, and I have not heard one positive story about them.

Chaintech makes a Mini-ATX board which isn't very mainstream.

This means, that as of now, the K75SA = SiS 735 to the buying public. As for the problems exhibited, here are some explanations:

1) In most cases, a Thunderbird 1.4 could not run under stable conditions without the addition of a capacitor. This is totally unacceptable.

2) The memory issue can be seen by running a program that I think is called MemTest86 (? - can't remember the exact name). It interfaces directly with the memory, circumventing the OS, and I have never seen a 735 board not have several errors during this test.

3) The sound issue is something I have seen in my own experiences, and has to do with Audigy drivers not working correctly on a fresh install of WinME on a SiS735 board.

4) For those that think the SiS AGP Drivers don't have issues, try to run a Radeon 7500. They have offered no fix or solution, yet their is not one graphics card that is totally incompatible with VIA chipsets (aside from early AGP issues with Socket 7 and Geforce cards which was later resolved).

 

[Duvie]

< Leadtek has a whole new host of issues, and I have not heard one positive story about them. >

This proves my point...If it was a sis chipset issue the problems would be the same on all boards with the sis735...You are not making a good argument...

< 1) In most cases, a Thunderbird 1.4 could not run under stable conditions without the addition of a capacitor. This is totally unacceptable. >

Most of the ppl that had issues with 1.4tbirds here in the early stages were resolved with better power supplies. There was a rash of newbies buying this board cause it was cheap and expecting a 1.4tbird to work on in most case 300watts or less or non-approved amd power supplies. Once again I think your most claim is inaccurate.

< 2) The memory issue can be seen by running a program that I think is called MemTest86 (? - can't remember the exact name). It interfaces directly with the memory, circumventing the OS, and I have never seen a 735 board not have several errors during this test. >

You got the program named correctly. I have ran it with no errors and there was a post recently where many stated they got no errors. Contact insane3d and see if he got errors with the leadtek. I beleive his issues with memory was compatability issue of that board with crucial and the perceived ocing benefit do to ddr voltage adjustments left him disappointed.

3) The sound issue is something I have seen in my own experiences, and has to do with Audigy drivers not working correctly on a fresh install of WinME on a SiS735 board.

That is funny I had no such issue with winme and audigy.

4) For those that think the SiS AGP Drivers don't have issues, try to run a Radeon 7500. They have offered no fix or solution, yet their is not one graphics card that is totally incompatible with VIA chipsets (aside from early AGP issues with Socket 7 and Geforce cards which was later resolved).

I have not heard of this so it may be true...The fact it only occurs with the 7500 and not with 8500 (since pabster runs 8500's) though it is the same product line makes me really question this.

 

[AZGamer]

I have the following two links to show you for the problems I described. Would you care to explain why this isn't a chipset issue?


Data corruption issue

Also, the Radeon 7500 is based on the old Radeon core, whilst the Radeon 8500 is based on the new RV20.

Also, when I said Leadtek brought in new issues, I meant issues in addition to those that ECS boards face.

 

[Duvie]

"It appears that boards with 0, 1, or 2 (maybe also 3) on the sticker by the PCI slots usually have 40 Ohm ZP and ZN resistors on the board. Boards with 4+ usually have 40 Ohms for ZN and 56 Ohms for ZP. This change seems to approximately coincide with the approval for Athlon XP processors. Although the reasons are not known, an AMD document has been circulating that states motherboards designed for XP processors should use 56 Ohms for ZP as a starting resistance. "

Man read your own f^ckin links....It is a motherboard manufacturers problem, hence the board revisions. They did not follow amd specs and made a mistake...There has been no revised sis chipset...This is ECS's fault and problem.

Also some have disputed the claims made by mrathlon...I think he may have something but you still fail to show this is a chipset issue.

Your other links I can't read...I guess I am not a subscriber...

There has been one agp update....If the 7500 is running off of older gpu then whty are the older radeon cards not having issues???

 

[AZGamer]

Here is a fixed link for the Radeon 7500 issue:

Radeon 7500 issue

 

[Duvie]

Once again I fail to see you prove a point...Do you read this stuff? No one suggest its the chipset or even the agp driver....ONe radeon 7500 owner running same oS reports no issues...

I like this quote...I think it sums up most conditions we read in tech support forums...

<_I'm_ not having any problems. At all. I went from a stable, but aging, BX system to this K7S5A, overclocked the hell out of it, and it's still MORE stable than my BX ever even tried to be. And a LOT of people aren't having any problems. Remember, the majority of posts WILL BE ABOUT PROBLEMS, because when something works, we tend to take it for granted and don't bother telling other people >



I also found this and thought it goes with above...take it for what it is heresay, but since I don't frequent ocworkbench I have to take this as the state of ecs issues at ocworkbench.com...I noticed that the 18page thread above had realtive small percentage of new post in the last 2-3 weeks...

This is in response to someone suggesting it may be related to memtest errors of your other link, whixh I think I showed clearly is not a sis related item...

< If you have a look at the ECS section of the OCWorkbench forum, you'll note that more and more people are posting that they've cleared up their problems by replacing their power supplies with ones that are of higher quality or provide more voltage on the +3.3 and +5 rails. >

 

[AZGamer]

Let my clarify this. I am not a VIA-zealot, but I also don't like the misinformation spread around about the stability and issues with VIA and competing chipsets.

I have successfuly debunked the myth that there are no issues with the SiS chipset - on the other hand, there are as many issues, if not more, then there are with VIA chipsets.

 

[AGodspeed]

This doesn't help the people with AMD processors, but take the Intel 815, 845, 850 chipsets. Where are all the bugs and flaws that "every chipset has". Those boards are truly rock solid. Use any soundcard you like, any video card you like.

Actually Oldfart, I'm gonna have to disagree with you here. Not all chipsets are perfect, and this includes i845/i850 boards. You should note that, quote:

Because of the lower voltage signalling, the AGP4X slot on i845 and i850 motherboards is notched so that 3.3Volt only cards cannot be inserted. A problem arises where older AGP cards exist that have the correct notch, but do not comply with 1.5Volt signalling, asking for 3.3Volts from the motherboard instead. In such cases, the video card and the motherboard can be permanently damaged.

Also note:

A number of cards have been identified that have the 1.5 Volt notch, but are actually 3.3Volt cards.

These are:
- some nVidia Riva TNT2 cards
- all nVidia Vanta, Vanta LT cards
- all SiS 6326, SiS 305 cards
- all S3 Savage4 cards up to, but not including revision 3.0
- all 3dfx AGP Voodoo 3, 4, and 5 boards (officially non-AGP 4x anyway)
- some nVidia GeForce2 GTS/Pro.

I'd have to do a search for the i815, but I'd bet it's not, as you say, "truly rock solid", as all chipsets are not.

 

[Duvie]

You haven't proved or in your words "debunked" anything...Every one of your arguments has been shot down and shown it is not a sis issue with the 735 chipset but the fact that ecs made some flaws on their end in design of the board.

The whole point of this thread was never to tout ecs mobo and sis735 as superior but to point out the real need and hope that the 745 chipset can solidify stability for amd and their platfroms. I ham hoping that more mainboard manufacturers will be putting out sis735 and 745 mobos. Asus doesn't make ohm resistor errors like ecs. I think a couple of the big boys can really put out a great sis735 board.


(clip from the opening post of this thread by me.)
< I hope sis can gain more of a foothold with other motherboard manufacturers and get some other 735 systems out as well. ECS I can admit seems to have a little quality issues such as the high number of boards shipped with failing cmos batteries and motherboards shipped with jumpers set incorrectly. Otherwise I think it is obvious the sis chip and its integrated north and south has performed quite well. I have so far not had to update any motherboard drivers unlike vias 4 in 1 drivers that are in a constant stream. >

This wasn't a via bashing as I am stating documented issues, not just finding a thread by someone with a problem and claiming it is a reflection of all the boards or the chipset.