Flaws in Via chipsets hit ATA/133, SCSI performance

[Jeff H]

Guys, why isn't this "issue" getting more attention? It sure seems to be real, w/ the article explaining why Via chipset-based boards (Intel and AMD cpu-based) have poorer hard drive performance than Intel, ALI, or SiS chipset based boards. At the end of the article you can link to the English version of the website that did the testing.

 

[Elledan]

"Flaws in Via chipsets hit ATA/133, SCSI performance"

 

[Gstanfor]

Oh, it's getting plenty of attention don't worry - here and at ARS-Technica.

VIA's PCI performance is utterly pitiful - something I have known since I "upgraded" my old Super-7 system from the ALI-5 chipset to VIA MVP3 and watched my video capturing capability (which was amazingly good for Super7) basically exit stage left along with HDD speed (for those wondering I use a Genius Video Wonder Pro2 BT848 PCI card for capture).

At the time I thought it was Super-7 related only. But the situation didn't improve when I upgraded to an Athlon (T-Bird) and KT133, nor did it improve when I upgraded the motherboard to KT133A. The KT133A is (very) likely to be the last VIA chipset motherboard I will ever purchase.

Greg

 

[Jeff H]

Ah, a rookie mistake on my part, not doing a search first! What got me to post this in the first place was the mind games I'm going through trying to decide between an Athlon XP or P4 system. For quite a while I was firmly entrenched in the Athlon camp, but now w/ the SiS 645 chipset I'm leaning toward P4. I'm going to be replacing a BH6/P3-650 system (it's going to college w/ my daughter next fall). It works fine for me right now, but I have the itch to build another machine, having not done that for myself in a couple of years. I've built more than a few recently, but for relatives or friends.

So now it's my turn, and I find myself impressed by the Asus P4S333 and the Soyo P4S Dragon boards. I certainly don't want to start an AMD vs. Intel flame war here. What I do find interesting is the pace at which things change, making my decision a very fluid one.

 

[Pabster]

Because there are so many VIA lovers here, that's why

It is a serious issue.

 

[Mem]

With Pabster around he won`t let us forget even though I`ve no problem with my Epox 8KHA+ .

 

[ToBeMe]

It IS an issue.........I was on the oppisite side too at first........but.......if you have high performance drives, B/M them and take a look at your results! I can use the same drives in non-Via setups of Intel or AMD and the results are pitiful in comparrison! There IS a problem, now, whether Via addresses it in a timely efficient manner or not remains to be seen.........personally, my bet would be that they will leave all the KT266A owners screwed and attempt a fix in the soon-to-be KT333 chipsets!

 

[Duvie]

I have recently b/m'd my drives in my newer sis735 mobo and hdtach and sisoft sandra scores are much much better....bootup is a tad faster, right around 20 seconds....

I am getting ready to add pci controller card so I can give each device its own channel and I remember how my scores were usually worse on a controller card in my old kt133 system...

I can only speak to 686a southbridge as I have no real b/m experience with the newer kt266 or kt266a boards...


The SB issue I cam overlook as I can avoid a specific piece of hardware, but this is much more serious as it covers the whole field...

 

[Duvie]

Well whether any of that is true or not, and you certainly have the right to your own opinion (eventhough I disagree), you shouldn't have to tweak anything to get standard performance out of ones components they buy. Especially when other boards and chipsets (intel and amd) give it to you.

Think about it...via boards (kt266a) are some of the most expensive chipsets and they seem to be the ones with the greatest issues. ECS quality control??? I agree they are not the best, but I get the performance out of my drives, especially now that I am installing an pci ata100 controller...I shudder to think the performnace hit I may have taken.

 

[MadRat]

The only flaw I've seen with Breeze's "memory enabler" is how Norton's Anti-Virus doesn't like it unless you make the NAV start upon Windows startup rather than with DOS startup.