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

[VBboy]

(Quoted)

A serious design defect in Via chipsets results in boards based on them substantially underforming motherboards with chipsets from Intel, SiS and ALi, a series of tests conducted by tecChannel.de has shown. The problem affects boards using both Intel and AMD chips, and the hit to hard disk performance is sufficient for tecChannel to say: "we can currently not recommend VIA chipsets for professional users who demand high performance from their hard drives and think about setting up RAID configurations."


http://theregister.co.uk/content/3/23502.html

What do you think?

 

[LostHiWay]

repost, but a very serious problem. Even with the so called patch by VIA which doesn't really work

 

[Viper GTS]

What do I think?

I think I'm damn glad there aren't any VIA chipsets in my system.



Viper GTS

 

[VBboy]

Sorry, didn't see it posted before.. Should I ask for this to be locked? Maybe keep it open for those who haven't seen the other post..

Please let me know if this post upsets anyone

 

[ToBeMe]

<< Sorry, didn't see it posted before.. Should I ask for this to be locked? Maybe keep it open for those who haven't seen the other post..

Please let me know if this post upsets anyone >>


Yes..........it upsets me.........:|

BUT........not because of it being a double post.................It upsets me because, I have a couple of via chipset boards in use right now and have built a few around Via chipsets lately too......hopefully this gets worked out by a patch/update...............

 

[uhohibrokeit]

its a big problem, but most likely a driver issue... glad im not using faster than ATA100 on a VIA board right now anyways

 

[IgoByte]

This is terrible. I actually hadn't heard about it before. Man, am I glad I got that Tiger MP...

 

[bacillus]

me smiles with my intel chipset m/board!

 

[oldfart]

<< me smiles with my intel chipset m/board >>

 

[randomlinh]

blah, stupid via :disgust:

 

[Flatline]

Well, you could always use SiS...

 

[JellyBaby]

This is why Via gets the bad rap.

 

[VBboy]

This really sucks, by the way. I spent a lot of time researching Via KT266-A based motherboards, and finally decided to get the Soltek SL-75DRV2 - very fast and stable... Hmm. Not anymore?

 

[Yoshitoshi]

Yes, but this is concerning IDE and SCSI controllers as PCI cards, as the issue is with the PCI bus. Anyone read anything about issues with on-board controllers on VIA boards?

My KT7A-RAID does the job for me and I have few complaints.

Yoshi.

 

[NicColt]

Ok, I've been following this for a while but isn't this being overblown a little.

In these TBreak Shuttle AK35GTR articles and other article it compares the raid performance with an IWill 333-R Ali Magik chipset and an MSI K7N420 nForce Chipset. and I'm not even sure if he patched the VIA chipset from this and it still beats the nForce and the Ali Magik chipset in the SiSoft drive bench ?

Even the Abit KR7A Non-Raid was faster than the nForce or Ali Magik boards. Ok so there's a problem with the VIA chipset, I'm sure that it can be fixed with a future driver update. But is it so critical that it would make you rethink what board you will get. Am I missing something here. Maybe this thing doesnt affect onboard RAID just PCI raid cards ?

 

[Diable]

It just effects PCI cards.

 

[Doomguy]

The issue won't affect you unless youy're running at least 3 drives in RAID. Use some common sense, most drives don't come close to utilizing ATA/66.

 

[nightowl]

This issue goes farther than just PCI cards, it affects the entire PCI bus. With the enitre PCI bus being limited the overall system performance can be slowed when the bus becomes saturated. So, your PCI NIC, Sound Card, IDE controllers (both onboard and add-in, they are both connected via the PCI bus), SCSI, PCI graphics all contribute to the total amount of bandwith of the PCI bus. If the entire bus is bandwidth limited compared to other chipsets then when under a full load the system performance will suffer.

 

[Rahminator]

No, this issue doesn't affect onboard IDE controllers with new VT8233 southbridges but it does affect 686b. This bug affects only devices hooked up to white PCI slots and RAID, even if it's onboard.

 

[VBboy]

Doomguy - << The issue won't affect you unless youy're running at least 3 drives in RAID. Use some common sense, most drives don't come close to utilizing ATA/66. >>

Not really - the burst speed is when data is transferred to/from the Drive's interface, not its surface. So if something is cached in the drive, it can easily reach the theoretical limit of the interface, e.g. 100 or 133 MB/sec. Of course this could only go on for a very short period of time, but it still matters.. But when duming a large file from memory to disk, it will of course be limited by the physical throughput of the drive, e.g. 20 MB/sec or whatever the drive is capable off..

 

[Doomguy]

Sorry but burst speed has little to no effect on performance.

 

[NicColt]

>No, this issue doesn't affect onboard IDE controllers with new VT8233 southbridges

This Shuttle AK35GTR is looking very good then

[Mr. Burns] Excellent..... MUAHAHahahaha [/Mr. Burns]

 

[Pabster]

Ah, I see the excuse makers have come out

This issue has been "reported" for quite some time now. It's only recently been "confirmed" to affect KT266A/8233 as well.

"But it only affects PCI add-in controllers..."

Not true. It affects data transfer rates and processor utilization from both on-board IDE devices, as well as PCI add-in controllers.

It's no surprise, really. Each and every VIA chipset has suffered from laughably low IDE performance. It's pretty disappointing when the lackluster ALi chips even surpass VIA in these areas.

 

[Damascus]

<< Doomguy - << The issue won't affect you unless youy're running at least 3 drives in RAID. Use some common sense, most drives don't come close to utilizing ATA/66. >>

Not really - the burst speed is when data is transferred to/from the Drive's interface, not its surface. So if something is cached in the drive, it can easily reach the theoretical limit of the interface, e.g. 100 or 133 MB/sec. Of course this could only go on for a very short period of time, but it still matters.. But when duming a large file from memory to disk, it will of course be limited by the physical throughput of the drive, e.g. 20 MB/sec or whatever the drive is capable off.. >>



Dude, largest cache on most IDE drives is like 2MB.

 

[VBboy]

Damascus,

uhh no.

WD has a special-edition HD with 8-MB cache, and we're also talking about SCSI. Try 16 MB+ there