Sign in with
Sign up | Sign in
Your question

this is why i hate Via chipsets!

Last response: in Motherboards
Share
January 2, 2002 3:16:36 PM

Flaws in Via chipsets hit ATA/133, SCSI performance
By John Lettice
Posted: 26/12/2001 at 08:46 GMT

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."

The problem is that Via chipsets are currently unable to take full advantage of the performance of PCI. Controller cards with Ultra-ATA/ 133 chips have a maximum theoretical throughput of 133Mbytes/s (or 127.2 in real megs), and tests of non-Via motherboards with Promise and Highpoint controllers (using either Maxtor D540X or D740X) showed burst mode transfer speeds of 95-117Mbytes/s. The same tests performed on a variety of Via boards came up with speeds of 63-78Mbytes/s, clearly indicating that there's something unpleasant going on here.

Forther investigation provided an explanation. According to tecChannel: "Normally, a burst should be performed continuously and without any interruption... The length of a burst, however, can differ depending on the data transferred.

"With VIA boards, this high speed transfer from the cache of a hard drive is constantly interrupted within a couple of µs. It has then to be re-initiated... Therefore, the effective burst rate drops to 64 to 90 MBytes/s at best... In contrast to chipsets from Intel, SiS and ALi VIA's products seem to have difficulties with maintaining high transfer rates close to the maximum speed of PCI for a longer time. As with ATA/133 PCI has a theoretical maximum bandwidth of 127,2. It seems fair to conclude that VIAs implementation of a PCI bus must be the reason for the problems found."

Via also underperforms when it comes to SCSI. Systems set up with Intel 845 and Via P4X266A, plus an Adaptec RAID-2100S, showed performance around 30 per cent better for the Intel system. Longer bursts used by the Intel system result in much better performance. So if you buy a high performance controller then put it in a board via a Via chipset, right now you're probably wasting your money.

According to tecChannel Via has confirmed the problem, and says it is working on a solution. However it is not clear whether new drivers will be enough, or whether it will need a redesign of the chipsets. There is an unofficial patch produced by George Breese of Networking Resources available here, but as tecChannel says, you use this at your own risk.

The full, highly detailed report can be read here. ®

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

<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>

More about : hate chipsets

January 2, 2002 3:41:25 PM

That's why my next mobo will be VIA-free!!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 2, 2002 4:00:16 PM

yep...

either AMD chip has a decent chipset or i'm going intel simple as that. Sis is a joke, via sucks, nforce isn't even out. ya ok like what; one board has it? And AMD doesn't put there chipsets into mass production/design.

my board is hybrid so i'm half via free heh. I figure one less via chip on my board is better then having both of them lol.

so far my next upgrade will be an Intel. If i do upgrade. I do like my current rig as it is. I may not upgrade for awhile. When 4ghz chips are out thats when i may upgrade.

AMD chips are nice. However where are the chipsets for it? the very thing that drives the cpu. You would think AMD would mass produce and design it to make there product look good. AMD made a mistake there.

<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>
Related resources
January 2, 2002 4:04:48 PM

nForce has been out for a while now! SiS is pretty good but we need something other than ECS! ALi Magik1 revision C is currently the best chipset for an AMD platform. Just look at the IWill XP333, it's amazing!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 2, 2002 4:55:27 PM

VIA chipsets have had IDE problems of one kind or another starting with the 686A southbridge. These problems were well documented by Tom about 2-years ago!

IMHO, nForce is a disappointment - especially considering the recently revealed DIMM-slot restrictions/problems.

SiS735 is a very nice integrated chip, but for reasons I just can't understand, none of the reputable board mfrs want to use it! I can only hope that situation will change with the newly released SiS745.

ALi MAGiK 1 (C1), is a stable, but not terribly exciting, performer.

<b>WISHING YOU ALL A GREAT 2002 !<b>

<b>God bless the <font color=red>U</font color=red><font color=white>S</font color=white><font color=blue>A</font color=blue></b>
January 2, 2002 5:54:15 PM

"...IMHO, nForce is a disappointment - especially considering the recently revealed DIMM-slot restrictions/problems...."

Were you planning on utilizing more than 1GB?

I love my nForce! No disappointment here. (Except for the printed manual, and lack of POWER LED headers - this is MSI fault).

<i>I love my nForce ... don't you?</i>
January 2, 2002 6:00:12 PM

nForce has been out on several boards for quite a while now. You should check PriceWatch before saying something isn't out.

The SiS735 has proven itself to be a hell of a chip (not chipset, I suppose), and the 645 looks pretty good, too.


BTW, a friend of mine won your board (Gigabyte AMD761) from AMD in their little roadshow. He's been VERY impressed with it. Most stable board he's ever owned, he says. Hasn't crashed once on him.

<font color=orange>Quarter</font color=orange> <font color=blue>Pounder</font color=blue> <font color=orange>Inside</font color=orange>
January 2, 2002 6:20:10 PM

well i have my vendors i shop for. newegg, mwave, and buy.com. I may see one bored that has the nforce chipset and i haven't heard much of anything about it.

as for the gigabyte board. I have had no problems with it whatsoever. Maybe i lucked out. I have had no problems.

But thats cool your friend won that board and hasn't any problems with it. yet anyway.

There is one problem.

if you connect a second hard drive to the promise cotnroller it sees it as PIO 4, but still runs at udma 5. If you set to optimized defualts and enable audio it fixes that. so the bios programers are a little fuct up.

there is a mod bios you can try out... crudd i never saved the address... i'll try and find it later.. it has the full raid version and uses the 7dxr+ bios .. suppose to resolve a lot of issues with it.


<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>
January 3, 2002 6:24:39 AM

I too have gone down the road of driver city to find a way around the numerous SERIOUS bugs in the VIA chipset. I just bought an ABIT KG7-RAID thinking I was set for stability and speed. I thought it didn't use VIA, but to my astonishment, I find a AMD 761 and a VIA 686B chipset!!! Argh! Of course it is about as stable as Clinton in a College Co-ed dorm.

Rob.
January 3, 2002 10:56:45 AM

Odd, after having 3 different VIA boards, I'm still waiting to have a problem related to the chipset. Hrmmm, maybe all you folks are just getting duds or are creating your own problems? (That's what I did, amazing how bad something can be screwed by a simple click)

Drivin' in a winter wonderland

Martin, MI
January 3, 2002 11:19:42 AM

Where have you been? There are tons of VIA problems! Haven't you read about them on the web? The ALi MAGIK1(revision c) is the most exciting chip! It is capable of running at 166MHz or 200 MHz (333MHz or 400MHz effective) without overclocking the PCI and AGP bus!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 3, 2002 12:18:46 PM

doh sorry to here it's unstable (lol)

ya so far i haven't had much of a problem with my hybrid. I have the gigabyte 7dxr 1.0 rev ... and it's been very stable. Just think it's a lot more stable then if you had a ful via chipset. motherboards are relativly cheap just buy another one down the road. I'd wait until USB 2.0 and maybe even built in firewire :)  .. onboard SCSI 320 controller and raid (raid is actually level 5 raid, raid 0 isn't raid at all but striping which only improves performance. when some says raid, level 5 raid is usually assumed)

so ya lol ... just my wish list of what i'd like to see on my motherboard along with everything else.

<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>
January 3, 2002 12:23:17 PM

Chill, dude. I'm not saying it doesn't have problems. I'm saying I haven't had any. I had an ATI AIW Radeon that compatibility issues with the 133, but THAT is ATI's fault and the card sucked anyway. My point is; I think a lot of problems that people blame on chipsets (and CPU'sm for that matter) are actually the result of other faulty hardware, poor installation, and other user-related screw-ups. There are an awful lot of VIA chipsets out there being used all over the place. If VIA was as bad as some people like to make them out to be, they wouldn't be as successful as they are. I'm not going to get a flamewar started, so I just going to leave it at that.

Drivin' in a winter wonderland

Martin, MI
January 3, 2002 12:33:44 PM

ya that is pretty exciting!

400mhz fsb. Rivals Intels bus speed! Lets hope AMD makes the correction for there next chip to utilize the 400mhz fsb. If they were smart they would do more then just a die shrink.

<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>
January 3, 2002 1:19:22 PM

VIA is popular because they make fast cheap chipsets! But VIA is not known for the best stability!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 3, 2002 1:20:29 PM

Yeah,that would be amazing! A 400MHz FSB Athlon XP would kill a Northwood clocked 500MHz more!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 3, 2002 4:10:14 PM

Quote:
It is capable of running at 166MHz or 200 MHz (333MHz or 400MHz effective) without overclocking the PCI and AGP bus!



OK, correct me if I'm wrong. Now assuming your talking about an Athlon. And assuming your talking about overclocking the Athlon bus to run that high. Now if you do that I heard you are overclocking the PCI and AGP as well.

Complicated Nit Picker
January 3, 2002 4:22:19 PM

Quote:


OK, correct me if I'm wrong. Now assuming your talking about an Athlon. And assuming your talking about overclocking the Athlon bus to run that high. Now if you do that I heard you are overclocking the PCI and AGP as well.

Ok, you are wrong! hehe With the ALi Magik1 (revision C) chipset in the Iwill XP333, you can raise the FSB without overclocking anything else because it supports 1:5 PCI divider and 1:6 PCI divider. In fact, if you unlock your processor, you won't even need to overclock the processor! All you need is RAM that can handle 333 or 400MHz!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
Anonymous
a b V Motherboard
January 3, 2002 4:37:52 PM

Sounds interesting, you bet. Pricewatch has no ALi based boards. Where can I see these? Are they for sale? Who is making them?
January 3, 2002 7:32:20 PM

I'm wrong. You weren't talking about an Athlon? You weren't talking about overclocking? Aren't
you saying you can raise the FSB to 166 or more and not raise the PCI or AGP at all, not even 1MHz. Your saying 0?


Complicated Nit Picker
January 3, 2002 7:42:13 PM

Let's take the Abit TH7-II for an example. That board can "lock" the PCI and AGP busses so that they don't move at all, even though the Front Side Bus (FSB) of your system is moving up (or down, conversely) by 1 MHz increments. I'm not sure exactly how that's done (by dividers, like AMD_Man was talking about), or by independant clock generators (unlikely). But it is an amazing feature. Perhaps Raystonn and/or FUGGER can comment more on this board in particular?

Anyway, you are partially correct. Raising the FSB does indeed raise the clock speed on the PCI and AGP busses (and ISA), on older motherboards. Newer boards are utilizing dividers so that overclocking is easier. Nobody wants to stress their video card (or even sound card) so much out of spec that they get strange errors.

I'm rambling, I forgot what the original question was :) 

<font color=orange>Quarter</font color=orange> <font color=blue>Pounder</font color=blue> <font color=orange>Inside</font color=orange>
January 3, 2002 8:02:40 PM

The only one I know of that uses revision C of the ALi Magik is the IWill XP333, search for that.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 3, 2002 8:03:06 PM

The only one I know of that uses revision C of the ALi Magik is the IWill XP333, search for that.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 3, 2002 8:03:20 PM

OK, your agreeing with AMD-Man and saying with these dividers the PCI and AGP don't change one bit. Technically not even 1 MHz. Because I wasn't saying it raised it a lot, but technical that it did raise it.

Complicated Nit Picker
January 3, 2002 8:07:31 PM

Quote:

I'm wrong. You weren't talking about an Athlon? You weren't talking about overclocking? Aren't
you saying you can raise the FSB to 166 or more and not raise the PCI or AGP at all, not even 1MHz. Your saying 0?


No, not exactly overclocking at all (unless your RAM isn't speced at 333MHz or 400MHz). With the IWill XP333, you can set it so that the PCI speed is a 1/5 or 1/6 fraction of the FSB speed. 166.6666666666MHz div 5 is 33.33333, 200/6 is 33.3333. At those speeds the AGP and PCI bus isn't overclocked at all! O MHz overclock!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 3, 2002 8:08:09 PM

Quote:

I'm wrong. You weren't talking about an Athlon? You weren't talking about overclocking? Aren't
you saying you can raise the FSB to 166 or more and not raise the PCI or AGP at all, not even 1MHz. Your saying 0?


No, not exactly overclocking at all (unless your RAM isn't speced at 333MHz or 400MHz). With the IWill XP333, you can set it so that the PCI speed is a 1/5 or 1/6 fraction of the FSB speed. 166.6666666666MHz div 5 is 33.33333, 200/6 is 33.3333. At those speeds the AGP and PCI bus isn't overclocked at all! O MHz overclock!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 3, 2002 8:08:30 PM

Quote:

I'm wrong. You weren't talking about an Athlon? You weren't talking about overclocking? Aren't
you saying you can raise the FSB to 166 or more and not raise the PCI or AGP at all, not even 1MHz. Your saying 0?


No, not exactly overclocking at all (unless your RAM isn't speced at 333MHz or 400MHz). With the IWill XP333, you can set it so that the PCI speed is a 1/5 or 1/6 fraction of the FSB speed. 166.6666666666MHz div 5 is 33.33333, 200/6 is 33.3333. At those speeds the AGP and PCI bus isn't overclocked at all! O MHz overclock!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 3, 2002 8:08:51 PM

Quote:

I'm wrong. You weren't talking about an Athlon? You weren't talking about overclocking? Aren't
you saying you can raise the FSB to 166 or more and not raise the PCI or AGP at all, not even 1MHz. Your saying 0?


No, not exactly overclocking at all (unless your RAM isn't speced at 333MHz or 400MHz). With the IWill XP333, you can set it so that the PCI speed is a 1/5 or 1/6 fraction of the FSB speed. 166.6666666666MHz div 5 is 33.33333, 200/6 is 33.3333. At those speeds the AGP and PCI bus isn't overclocked at all! O MHz overclock!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
Anonymous
a b V Motherboard
January 3, 2002 8:17:01 PM

Not only does the VIA chipset suck when it comes to ATA133 and SCSI, but it is not compatible with Creative Labs Sound Blaster. A month ago I tried out the SOYO Dragon Plus with Sound Blaster Live 5.1, and everytime it would boot to Windows, it would just lock up. I am currently using the IWill XP333-R with the ALi Magik chipset and it is great, I have had no problems with it so far, I recommend this motherboard to anyone.
January 3, 2002 8:22:32 PM

Your fingers must be sore writting that out 4 times. hehe I'll have to check up on this board. I understand they had some ups and down getting the board ready.

Complicated Nit Picker
January 3, 2002 8:51:38 PM

I see what you talking about now. I seen a few reviews and they were in conflick.

Complicated Nit Picker
January 3, 2002 10:55:04 PM

Hmm, how did that happen? lol THGC has been acting up lately!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
Anonymous
a b V Motherboard
January 3, 2002 11:07:16 PM

Thanks AMD_Man I will. Pun intended.
Anonymous
a b V Motherboard
January 3, 2002 11:17:29 PM

So.. Whats the deal with this? I bought my GA-7DXR (AMD 761 north, VIA 686B south) board about a couple months ago and just recently setup an array only to find out that.. what? A board BUILT with an INTEGRATED RAID adapter cant handle the data IO rate necessary for one? Does that make ANY sense to ANYONE? Did ANYONE bother to test this thing before the sold it to me??

So apperently some German guy comes up with a fix for VIA chipset boards, but not those like mine with a hybrid chipset. I guess Ill have to wait for some other European guy to make one for hybrid boards, I mean hell, VIA isnt exaclty on top of it.

But the part I like best is the Tech support guy at Giga-Byte. Before I read up on all this PCI Bus stuff with the VIA 686B, I described to him the trouble I was having. I think I mentioned corrupted data and illegal operations on my array of 2 BRAND NEW WESTERN DIGITAL drives that I told him I had tested independently of the array. He suggested I try different HDDs!!! WOW THANKS FOR THE ADVICE SIR!! Y00 Iz GuD TeKK Sp0RT!!

Ok Im done ranting. Thanks.
January 4, 2002 1:47:12 AM

It is AMAZING. PCI bus have been around for more then 5 years and VIA still cannot get the chipset work effiently.
January 4, 2002 11:08:52 AM

5 years? It's more like 10 years!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 4, 2002 12:04:56 PM

Sorry, have to disagree again. I have a Live! 5.1 and it works fine with my 266a. No popping, no crackling, no lockups or slowdowns. Am I just luckier than others or what's the deal? 'Course, I haven't tweaked it to get the absolute last smidgen of speed out of it either. I rely on this thing to work for business purposes and it hasn't let me down (unless I screw something up, of course) :wink:

Drivin' in a winter wonderland

Martin, MI
January 4, 2002 3:32:09 PM

On the Abit TH7-II, it does not change even 1 MHz. On other boards (with dividers), yes it changes.

<font color=orange>Quarter</font color=orange> <font color=blue>Pounder</font color=blue> <font color=orange>Inside</font color=orange>
Anonymous
a b V Motherboard
January 4, 2002 3:39:51 PM

hmm.... some of you guys have to do your homework... via fixed its problems with its 686a southbridge and released a 686b southbridge later that year.. tom did a full review of the chipset with the new benchmark results... vastly improved.....

i happen to own a abit vh6-2 mobo with the via 686b south and 694 chips and it runs without problems (bios update 8/13/2001). the hard disks i have are a 1 wd 10gig caviar at uta66, and 2 on a promise uta66 raid comtroller, even though i am not taking full advantage of the uta100 bus, i still have a full uninturupted burst from the disk cache. sandra 2001 pro reports that the buffered read was at 66mb/s from this disk and a buffered write of 63 mb/s on primary master on the ide bus. (on my system ... not benchmarks from toms)

however i have a promise fasttrak 66 in my first pci slot and i have excellent benchmarks from my other 2 10 gig caviars.

i have a buffered read of 102mb/s, a buffered write of 96mb/s. (from disk cache)

i would have to say this is pretty darn good... also... i have not had any problems or instability with this chipset... i'm running win98se and have 0 problems other than the normal ms crap.

maybe you guys should read or maybe even look at more recent hardware other than the stuff you have from 3 years ago... it helps... in fact... i would even recommend the via chipset (on an abit board) to someone... very reliable and i have excellent performance...

MS.crap has performed an illegal operation in run.suck. If the problem persists reboot to Red Hat.
January 4, 2002 3:50:39 PM

apparently you can't READ! Let me post it again. Take a reading class or something and read the following.

"Flaws in Via chipsets hit ATA/133, SCSI performance
By John Lettice
Posted: 26/12/2001 at 08:46 GMT

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."

The problem is that Via chipsets are currently unable to take full advantage of the performance of PCI. Controller cards with Ultra-ATA/ 133 chips have a maximum theoretical throughput of 133Mbytes/s (or 127.2 in real megs), and tests of non-Via motherboards with Promise and Highpoint controllers (using either Maxtor D540X or D740X) showed burst mode transfer speeds of 95-117Mbytes/s. The same tests performed on a variety of Via boards came up with speeds of 63-78Mbytes/s, clearly indicating that there's something unpleasant going on here.

Forther investigation provided an explanation. According to tecChannel: "Normally, a burst should be performed continuously and without any interruption... The length of a burst, however, can differ depending on the data transferred.

"With VIA boards, this high speed transfer from the cache of a hard drive is constantly interrupted within a couple of µs. It has then to be re-initiated... Therefore, the effective burst rate drops to 64 to 90 MBytes/s at best... In contrast to chipsets from Intel, SiS and ALi VIA's products seem to have difficulties with maintaining high transfer rates close to the maximum speed of PCI for a longer time. As with ATA/133 PCI has a theoretical maximum bandwidth of 127,2. It seems fair to conclude that VIAs implementation of a PCI bus must be the reason for the problems found."

Via also underperforms when it comes to SCSI. Systems set up with Intel 845 and Via P4X266A, plus an Adaptec RAID-2100S, showed performance around 30 per cent better for the Intel system. Longer bursts used by the Intel system result in much better performance. So if you buy a high performance controller then put it in a board via a Via chipset, right now you're probably wasting your money.

According to tecChannel Via has confirmed the problem, and says it is working on a solution. However it is not clear whether new drivers will be enough, or whether it will need a redesign of the chipsets. There is an unofficial patch produced by George Breese of Networking Resources available here, but as tecChannel says, you use this at your own risk.

The full, highly detailed report can be read here. ®

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




<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>
Anonymous
a b V Motherboard
January 4, 2002 7:06:18 PM

Fellas, mellow out.

Yes, VIA has an issue with the timing of their burst timings when dealing with ATA/133, but it's not that big of a deal. Realistically, this only comes into play when ATA/133 PCI devices are used on the motherboard. ATA/100 devices like those built into the motherboard do not have this problem. Additionally, there is already a software fix for this that was generated by George Breese of Networking Resources which modifies the timings to allow much larger bursts of data. That's the flaw, VIA chipsets chop their bursts to frequently and cause them to have to be resent. With this simple fix, here's the results:

Chipset------ No Latency-Patch --- With Latency-Patch
VIA MVP3----- 63,5 MBytes/s ------ 84,1 MBytes/s
VIA KT133A--- 78,2 MBytes/s ------ 93,4 MBytes/s
VIA P4X266A-- 90,1 MBytes/s ------ 100,9 MBytes/s

These speeds are completly comparable to other chipsets.

Now I'm not saying that there isn't a problem because obviously there is. I'm just saying that it's not the OMFG I'M GOING TO DIE sized problem everyone is making it out to be. Clearly, an upcoming BIOS download will take this into account. However, my Asus A7V266-E is never going to need it because I'll never get up to those rates of traffic on my PCI bus. An I'm a pretty demanding. However, I don't need to have a 4 disk raid setup on my machine. So if you do, I guess you should reconsider a purchase of a VIA based board right now. For everyone else, who cares?

Ron
January 4, 2002 7:57:45 PM

From what I was reading the I will333 doesn't change either. At first I thought it did, it must of been old stuff I was reading. It sounds like the dividers are a good ideal and looks like things are going to be speeding up in the future.

Complicated Nit Picker
January 4, 2002 8:08:01 PM

I'm pretty lucky too I have the Live 5.1 Gamer.

And I'm also glad to see not everybody is a VIA hater. Accept for my AMD chip all my others been VIA and I have no problems, just the usual stuff. All the other chip makers had their problems too. That's life.

Complicated Nit Picker
January 4, 2002 8:16:05 PM

I agree with what you are saying to a degree but I wonder about efficacy of the PCI latency patch. I have a KT133A mobo and a Maxtor D740X and I am only seeing burst rates of 80 MB/sec on an ATA100 interface.

80 MB/Sec or 100 MB/sec doesn't really matter for a single IDE hard drive. Even at the slower speed, a hard drive's 2 MB cache is depleted/saturated in 0.025 seconds. After that transfers are limited to the sustainable speeds of the hard drive. In the case of the D740X, that's about 45 MB/sec Max and about 35 MB/sec average. This doesn't even meet the promise of ATA/66, not yet. It's even worse for writes.

Now, if you talking about taking SCSI or ATA/133 drives (the latter with 8 MB caches) and stripe four of them and force them to share the bandwidth of a PCI bus artificially slowed down to 80 MB/sec (or 100 MB/sec as you assert) then transfers are really being slowed down.

<b>We are all beta testers!</b>
Anonymous
a b V Motherboard
January 13, 2002 4:19:12 PM

The fact is that VIA chipsets may have a lot more bugs than that one. It's just that THIS TIME someone could actually PROVE that VIAs chipset is broken by design!

What about all the people with AGP problems, infinite loop in vn4_disp.dll etc. Try checking how many driver releases VIA has on their website (one every week?). Then go check out Intel. Very few fixes are needed!

VIA quality is dropping below acceptable standards. So who cares if they are sometimes first in a banchmark!!
January 13, 2002 4:35:28 PM

Quote:

VIA quality is dropping below acceptable standards. So who cares if they are sometimes first in a banchmark!!

Exactly my point in the Windows 95/98/ME/XP forum. Who cares if VIA wins benchmarks, if they have really buggy chipsets?

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
January 13, 2002 8:06:27 PM

intel uses VIA chipsets also?

please do not get confused between the chipset which controls the data flow and the CPU which does the processing. They are completely seperate and one does not cause the other.

There are lots of parts in your computer not just the intel cpu or the amd cpu, which are both good cpu's. Just there are other hardware in your system to then just the processor that can cause problems.

<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>
!