ALI vs. VIA

[agnusdei]

Ok everyone, time for some personal opinion. I want everyone who HAS OWNED and knows these chipsets to reply with what they prefer more. Don't give me your opinion on other chipsets, don't give me "they are both equally good" because we know thats not true. Please reply back and tell me which chipset you prefer. -Dave

 

[Pabster]

LOL

I'd take the ALi over the VIA. You'll sacrifice a little performance, but save yourself a hell of a lot of headaches. Just look at the number of VIA issues crowding any given discussion board

 

[Rand]

Not going to bother getting into a long winded discussion, I'll just state my preference.

VIA.

 

[Bahelion]

I have had both and each has their strong points. I will say the Ali.

 

[MadRat]

I side with Rand on this issue.

Just too many bad experiences with ALi and their Alladin sets.

 

[odog]

never like ALI.... thier last chipset was a steaming piece of sheet.....(A7A266)

VIA has improved.... and the KT266A is the fastest athlon DDR chipset out...

 

[AA0]

From my expirence, and from someone else who also owns an ALi and VIA chipsets, its VIA. ALi has major compatibility problems with a lot of his hardware, and some of mine, but no where near as much as his. He actually was forced to take apart his old system and mix the two to get them to work, Radeons perform awful on them, even with the patch.

 

[ST4RCUTTER]

Having owned SiS, VIA, and ALi, I can say my preference (right now) is still VIA.

Pabster says, "Just look at the number of VIA issues crowding any given discussion board"...he could equally say "Just look at the number of idiots crowding any given discussion board."

VIA is by far the most prevalent third party (not Intel) chipset maker in the industry so it shouldn't be surprising to see so many threads for VIA on these boards. Let's remember that virtually no one comes to these boards to post on how great their motherboard is running. They come when they're having trouble...which means the odds of an idiot getting a VIA board and posting about it far exceed those getting SiS or Ali.

People pine on and on about how many issues VIA boards have. Here are some of the major ones:

VIA 686B + SB Live = HDD data corruption, playback errors etc. (fixed by 4in1)
VIA + WinXP + Det 23.11 drivers = memory instability (fixed by 4in1)
VIA + PCI ATA133 controller cards = lowered IDE performance (may not be driver-fixable)

Beyond these there's not much else to cry about. The first two can be fixed, and the last one is fairly uncommon on the home desktop.
That doesn't mean that VIA shouldn't be taken to task for these ingressions, just that they're not as disastrous as some would have you believe. The performance of ALi is lackluster at best. Even their newer B0 stepping can't outperform the venerable AMD 760 chipset, whereas the KT266A sports the fastest DDR memory controller on the market. Some may argue that it's worth giving up performance "to save yourself a hell of a lot of headaches" but the reality is most headaches aren't from the chipset, but from the builder's actions.

The 735 is a worthy chipset that was adopted by vitually no one. ECS is a joke, I don't care how many boards they sell. One board (completely ignoring Chaintech's sad little 735 board) with no real overclocking options*, no RAID, an entry level BIOS, and one of the worst manuals I've ever read is not the freedom of choice I enjoy. I'm hoping that SiS will be able to take off with the 745 and power some of the names associated with quality like ASUS, Epox, Abit, MSI etc. They need to...Silicon Integrated Systems Corp. is carrying a heap of debt, and they need to stick around to keep a spur in backs of the other makers.

Here's a good link comparing the chipsets in question

 

[NFS4]

Nice post ST4RCUTTER, all the VIA-hating is getting quite old.

 

[Rand]

Well said ST4RCUTTER.
Well I'd agree VIA definitely has problems their way over exaggerated IMHO.

Personally, I have no issues at all trusting VIA to run my system with stability. The last few mobo's I've purchased have been VIA based, and I've been quite happy with VIA chipsets numerous times.

A slight addition to your brief list of major issues though:
VIA + nVidia DetonatorXP + WinXP= nv4_disp infinite loop bug.

Granted that's not a problem for everyone, but it is overwhelmingly frequent.
It's not totally VIA's fault though as it's more so an issue with nVidia's detonator drivers and occurs under many of chipsets albeit with less regularity,

 

[NFS4]

<< Well said ST4RCUTTER.
Well I'd agree VIA definitely has problems their way over exaggerated IMHO.

Personally, I have no issues at all trusting VIA to run my system with stability. The last few mobo's I've purchased have been VIA based, and I've been quite happy with VIA chipsets numerous times.

A slight addition to your brief list of major issues though:
VIA + nVidia DetonatorXP + WinXP= nv4_disp infinite loop bug.

Granted that's not a problem for everyone, but it is overwhelmingly frequent.
It's not totally VIA's fault though as it's more so an issue with nVidia's detonator drivers and occurs under many of chipsets albeit with less regularity, >>


The infinite loop is not just limited to VIA chipsets. I got the same errors with my GF3 and an ECS K7S5A.

 

[Rand]

<<

<< Well said ST4RCUTTER.
Well I'd agree VIA definitely has problems their way over exaggerated IMHO.

Personally, I have no issues at all trusting VIA to run my system with stability. The last few mobo's I've purchased have been VIA based, and I've been quite happy with VIA chipsets numerous times.

A slight addition to your brief list of major issues though:
VIA + nVidia DetonatorXP + WinXP= nv4_disp infinite loop bug.

Granted that's not a problem for everyone, but it is overwhelmingly frequent.
It's not totally VIA's fault though as it's more so an issue with nVidia's detonator drivers and occurs under many of chipsets albeit with less regularity >>


The infinite loop is not just limited to VIA chipsets. I got the same errors with my GF3 and an ECS K7S5A. >>



I know, see my original post "occurs under many of chipsets albeit with less regularity".
It's primarily seen with VIA chipsets, but is on occasion seen with other chipsets.

One good thing about VIA, at least they always try to fix their issues. One can't always say that about the cpmpetition. VIA's works hard to correct their issues and improve their chipsets. The 4in1 drivers are fequently updated.

It wasnt too long ago that we all said VIA memory controllers were always bad, and it was always the primary reason for VIA's consistently lacking performance.
No one could ever say that about the memory controller of the KT133A, KT266A, P4X266A.

 

[Pabster]

Why is it that the K7S5A is branded as "not overclockable"?

I'm running a Thunderbird 1.4GHz @ 1650 (150x11 w/ Golden Socket), an XP1800+ @ 1690 (147x11.5), and an XP1900+ @ 1764 (147x12). All 100% rock solid.

The myth that SiS 735 (or ECS) is "low class" is just that -- a myth. These $55 wonders are tops in the price/performance arena.

 

[Rand]

<< Why is it that the K7S5A is branded as "not overclockable"?
>>



It's overclockable, but I believe it's primarily stated as "non-overclockable" because the actual overclocking options are rather limited compared to that of most other boards that attain equivalent popularity with the enthusiast market.

 

[Pabster]

Ah yes, it doesn't have a slick little GUI bundled with it. Or a 10 page BIOS screen with a bunch of adjustments. I see

 

[Pabster]

ST4RCUTTER wrote:

"Pabster says, "Just look at the number of VIA issues crowding any given discussion board"...he could equally say "Just look at the number of idiots crowding any given discussion board."

As usual, the VIA defense begins.

The issues surrounding VIA chipsets are well documented. I'm not going to get in to another flame shoot over 'em, but to deny the issues you've already listed -- as well as several others -- or to make excuses for them -- is ridiculous. And the most ridiculous of them is that "VIA has a majority share of the market...". Of course they do. People see a few points in synthetic benchmarks and rush out to purchase VIA's latest and greatest ... only to be dismayed shortly thereafter as the usual crop of issues appear. You might not agree, but, there's certainly no shortage of people who would.

 

[HeSaidWhat]

I have used ALi and VIA and my preference is the AMD chipsets. They may not have all the speed of the VIA but they are the most stable board I have used.

 

[ShadowFox]

I have had 3 VIA boards and i have had NO PROBLEMS, PERIOD!

 

[ST4RCUTTER]

Why is it that the K7S5A is branded as "not overclockable"?

I'm running a Thunderbird 1.4GHz @ 1650 (150x11 w/ Golden Socket), an XP1800+ @ 1690 (147x11.5), and an XP1900+ @ 1764 (147x12). All 100% rock solid.

The myth that SiS 735 (or ECS) is "low class" is just that -- a myth. These $55 wonders are tops in the price/performance arena. --Pabster
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Fair enough. Let's list the overclocking options of the K7S5A:
-Multiplier adjustments of 100, 112, 133, 138, 150Mhz.


Now let's list the overclocking options of another budget board, Shuttle AK31 rev.3 ($80)
-Multiplier adjustments in increments of 1Mhz from 100Mhz-166Mhz
-Adjustable Vcore, DIMM, I/O voltages
-DRAM timings

Now let's look at the manual. The K7S5A sports a manual of 37 pages of voluminous information. The Shuttle reference squeaks by with 76 pages of text. The BIOS section of the ECS docket comes out to 12 pages...nice. The Shuttle manual gets this done in a "mere" 33 pages of reference. I don't know about you guys, but detailed manuals make my wallet feel better.

And I never excused VIA for offering products with compatibility problems...only that those problems aren't nearly as bad as you try to make them out to be.

 

[AA0]

Pabster, why do you ignore all the troubles other chipsets have? They all have issues, and they are far more major then VIA issues, if you ask me. You do nothing but show how biased you are.

 

[Pabster]

ST4RCUTTER:

K7S5A's ICF supports 100,112,124,133,138,142,146,150,160,166,200 FSB.

Who needs 1MHz increments?

Evidently, some people haven't done their homework. No surprise

Candoman:

As for the Golden Socket costing "as much as the board", you're wrong again. While PlyCon sells them for about $45-50, I picked 2 of them up locally for a mere $29 each.

Your fabled 8KHA+, assuming it actually POSTs, is lucky to hit high 140-150 with decent RAM. While there are exceptions, the 160ish+ which 8K7A was quite famous for just isn't realistic. I can do 150 as well with K7S5A, but I have to reduce RAM timings a bit. I could easily do the mod and crank the DIMMs to 2.7 or so, but why? For a couple extra MB/s of bandwidth? Nah.

 

[NFS4]

<< As for the Golden Socket costing "as much as the board", you're wrong again. While PlyCon sells them for about $45-50, I picked 2 of them up locally for a mere $29 each. >>


$55 + $29 = $84. And I don't think that everyone can "pick them up locally" for $29. So let's say $50 + $55 = $105...some bargain:| Even at $84 (your cost), I'm sorry, I'd rather have a Shuttle AK31 Rev 3.1 which has everything in BIOS and costs $4 less. Not to mention an added PCI slot, 2 more DDR DIMM slots, and 2 more fan headers. And the added CPU core voltage and DDR voltage tweaks.

 

[Pabster]

NFS4 wrote:

"$55 + $29 = $84. And I don't think that everyone can "pick them up locally" for $29. So let's say $50 + $55 = $105...some bargain Even at $84 (your cost), I'm sorry, I'd rather have a Shuttle AK31 Rev 3.1 which has everything in BIOS and costs $4 less. Not to mention an added PCI slot, 2 more DDR DIMM slots, and 2 more fan headers. And the added CPU core voltage and DDR voltage tweaks."

Actually, it was a better bargain on that particular board. I bought it for $35 from a trader who didn't know what he was doing and thought it had to be RMAed.

Anyway, pertaining to Shuttle's AK31A, I've used several. Never yet seen one that was stable with all 4 DIMMs populated, and that's at stock speeds. Also, I've yet to see one that actually could raise the core voltage. I always had to select one of the + options to keep it near 1.75, where it should have been by default. And I sure never got them anywhere near 150 FSB.

Fan headers are a moot point. All my fans go directly to the PS. I could care if the board had any or not, cause I'm not using them.

To each their own I'm just pointing out that this myth floating around that K7S5A "can't be overclocked" is baloney.

 

[ST4RCUTTER]

K7S5A's ICF supports 100,112,124,133,138,142,146,150,160,166,200 FSB.

Correct you are...for the most part. The ICF may support up to 200Mhz, but the BIOS will only support 100, 100/133, 112, 124, 133, 133/166, 138, 150, 166 settings (updated BIOS' not withstanding). According to Accelenation.com

Who needs 1MHz increments?

The guy who can't get his system to post at 140Mhz. He could try bumping up his Vcore to...oh wait, no he can't. You see my point.


Evidently, some people haven't done their homework. No surprise

Now you're just being petty.

 

[NFS4]

<< NFS4 wrote:

"$55 + $29 = $84. And I don't think that everyone can "pick them up locally" for $29. So let's say $50 + $55 = $105...some bargain Even at $84 (your cost), I'm sorry, I'd rather have a Shuttle AK31 Rev 3.1 which has everything in BIOS and costs $4 less. Not to mention an added PCI slot, 2 more DDR DIMM slots, and 2 more fan headers. And the added CPU core voltage and DDR voltage tweaks."

Actually, it was a better bargain on that particular board. I bought it for $35 from a trader who didn't know what he was doing and thought it had to be RMAed.

Anyway, pertaining to Shuttle's AK31A, I've used several. Never yet seen one that was stable with all 4 DIMMs populated, and that's at stock speeds. Also, I've yet to see one that actually could raise the core voltage. I always had to select one of the + options to keep it near 1.75, where it should have been by default. And I sure never got them anywhere near 150 FSB.

Fan headers are a moot point. All my fans go directly to the PS. I could care if the board had any or not, cause I'm not using them.

To each their own I'm just pointing out that this myth floating around that K7S5A "can't be overclocked" is baloney. >>


I like being able to check the fan speed on my fans which is a big plus. Also, I'm using 3 DIMM's just fine. If it isn't stable with 4 DIMM's, just up the DDR voltage to 2.6 or 2.7. It won't hurt anything. At least Shuttle gives you the option.

But for out of the box overclocking with voltage/bus/clock manipulation, the ECS K7S5A comes up way short.