3-Way Motherboard Shootout

Steve writes "Hexus.net has put three high-end i955X-based motherboards through their paces, to see which is the best LGA775 platform motherboard. Intel's own offering falls a little short, but Gigabyte and ABIT both make compelling boards, with ABIT taking the top-spot by a small margin."

Comments

No guns were found

[prurientknave]

The test ended rather abruptly when it was found the motherboards were not equipped with guns as ordered.


Move along. Nothing to see here folks.

Re:No guns were found

[Fastball]

The test ended rather abruptly when it was found the motherboards were not equipped with guns as ordered.

However, a lascivious motherboard menage a trois ensued when an enterprising hacker introduced the "Hot Silicon" mod. Senator Clinton is investigating.

three way

[DavidLeeRoth]

it would have been even more embarrassing if the article was called "Threesome Motherboard shootout."

Re:three way

[Adult film producer]

Thanks for the idea, always looking for something new and exciting.

Re:three way

[rainman_bc]

Odd, a friend of mine bought a Gigabyte board and had some serious issues with a bad production run.

After a couple weeks and lots of trouble with his machine, he found his board was missing a capacitor next to the processor that later versions had. No mention of the problem on Gigabyte's web site, some guy on a forum found it by comparing an earlier run with a later run to find the missing capacitor. They were good to repair it, but they covered it up pretty good too..

I'm not a huge fan of Gigabyte. Honestly, I've grown fond of Soltek... My last one really gave me no trouble at all, and I think I'd give them another chance, for a less expensive board they are really nice...

What about quality?

[CyricZ]

How is the quality of these boards? Will they still be working in two or three years? Or will they have leaking capacitors by that time?

Re:What about quality?

[WIAKywbfatw]

How is the quality of these boards? Will they still be working in two or three years? Or will they have leaking capacitors by that time?

You want to know if they'll still be working in two or three years time? Ok, let me just take a short trip in my time machine so I can find out.

Oh wait...

Re:What about quality?

[hab136]

The leaking capacitors were an industry-wide problem.. it was basically a crapshoot whether or not your board had them.

http://channels.lockergnome.com/news/archives/2005 0801_leaking_capacitors_muck_up_motherboards.phtml

Re:3-ways are hot.

[Ceribia]

"3-ways are hot" This is slashdot, thats not insightful it's redundant. Seriously, find one of us who didn't say it internally the moment we looked at the article.

I prefer the Gigabyte

[StarHeart]

Even if the Abit has a slight performance lead, I like all the features on the Gigabyte more. It is nice to see two pata connectors instead of just one like the other two boards. It also has more PCI slots than the Abit.

high end?

[vbrtrmn]

Wow, no SLi on any of them, I'll stick with my ASUS for now.

Re:high end?

[KillShill]

well thankfully SLi is a useless feature except for the most hardcore gfx enthusiasts.

i consider myself a gfx groupie but i wouldn't touch SLi with a 256ft pole. too much expense for too little return. gfx technology is moving too fast for SLI to become viable to more than a very small set of people.

you're also looking at extra cost in electricity in running 2 gfx cards. extra juice from the PSU is also required; meaning a more beefier unit which most people lack. it also takes up 2-4 slots total depending on the cards in question. it's a good option for the very few who have the money and willingness to go for it. otherwise, save up your money and in 6 months, you can buy a virtually similar performing card in a single slot solution.

if someone can find another reason to buy a motherboard with an extra high speed PCI-e slot... that's another story though.

Avoid Abit High Ends

[superpulpsicle]

I used to think brands really didn't matter. Until I bought the Abit Ka7-100. It was a highend board back around 2000 with 4 Dimm slots and 6 PCIs, 1 AGP. And an ISA slot which could be used when sacrificing the last PCI slot. It was fantastic at the time.

Then the transistors fried. I paid for shipping etc and got a replacement. Then it fried again, replaced, then fried again. 4 years later Abit sends me a letter saying they lost a lawsuit for selling select board models with broken transistors including the Ka7-100. Basically they knew it, and told consumers nothing about it.

Faster, higher, stronger?

[Anonymous Coward]

Yes, we get it. You guys can build big motherboards with gobs of features, including legacy stuff I'll never use.

For once, I'd like to see somebody build a small motherboard with modern connectors, quiet (or no) fans, and no legacy crap that just takes up space. Right now, about the only way to get that is out of a Mac mini.

I'm still using a 5-year-old computer that I've only upgraded once (bought a new 80 GB disk to replace the old 9 GB one). It's still fast enough for everything I do -- if/when I upgrade, I hope to get something smaller and quieter. I don't care about faster, because the slowest new computer available is faster than this sub-GHz Athlon.

It's kind of sad that, over the past 10 years, the only thing that's really changed about motherboards is that they've gotten faster. They've gone to all the work to come up with new and faster and smaller busses (Firewire, USB2, S-ATA, etc.), but it doesn't make the boards any smaller if they're still including all the big old legacy ports, still, too!

(Yay, floppy, IDE, and PCI slots! What, no ISA?)

ISTM that all motherboard manufacturers today build pretty much the same product, which is aimed at ... everybody, which means "nobody in particular". This is a market ripe for picking. If somebody had the brains and balls, they could really make a killing. How about a Mac mini-like motherboard?

Re:Faster, higher, stronger?

[eLDeR_MMHS]

There's been stuff out there for several years now.

A lot of the "small form factor" barebone cases by Shuttle, Biostar, Soltek, etc. might fit your spec.

Most of those cases are still at least four times bigger than a Mac Mini to accomodate full-size AGP/PCI-Express/PCI cards and 3.5" and 5.25" bays but then they are probably four times smaller than a regular PC tower. Price-wise they are somewhat more expensive than a DIY PC.

But yes, someone has thought of this.

Re:Faster, higher, stronger?

[Jeff DeMaagd]

I think you've somehow missed the chatter about the miniITX boards?

Re:Faster, higher, stronger?

[KillShill]

it's probably not a troll. just another user lacking insightful knowledge in the right context. it's not that legacy components are good but in his case he has no right to complain. a sub-ghz athlon, just upgraded his 9GB hd?

seems like by the time he gets around to upgrading his computer, Terabit ethernet would be long forgotten.

Re:Faster, higher, stronger?

[Velox_SwiftFox]

There are plenty of board to meet that request, most very small, inexpensive, and with soldered in CPUs.

3-Way ? Mother ? Bored?

[CDMA_Demo]

Is there a hidden message in this post that I cannot see?

Rewritten to fit title

[rietran]

Steve writes "Hexus.net has set three high-end i955X-based motherboards against each other, to see which is the best LGA775 platform motherboard. Intel's own offering fell after taking a shot to the head, but Gigabyte and ABIT both survived for hours, with ABIT taking out Gigabyte by dropping his gun and strangling him to death. Police are still investigating the cause of the shootout and Hexus' motive in provoking it."

pssst....

[KillShill]

these motherboards contain intel's DRM technology.

pass it on.

http://www.anandtech.com/printarticle.aspx?i=2449

search for DRM in the above site for confirmation and some extra info.

Re:pssst....

[Fastball]

pass it on.

I would but my motherboard prevents me from doing so.

Why can't they link directly to this?

[Mechcozmo]

http://www.hexus.net/content/reviews/review_print. php?dXJsX3Jldmlld19JRD0xNjAw

Printer friendly version for everyone so that this (click, load ads) doesn't (click, load ads) happen (click, load ads) to (click, load ads) you (click, load ads).

The most important thing...

[Anonymous Coward]

Only one thing matters. Warranty.
As a Genuine Intel Dealer, I can expect a replacement board to be overnighted even before I send the faulty one back. That way there is minimum disruption to the customer.
Try telling a customer their server is down for a few weeks while you wait for the board to be shipped back to Taiwan/China for testing before they'll issue a replacement.

Had a few customers who got non-Intel boards and had no computer for up to three months while waiting for a replacement. Think about that. That's three months paying for broadband you can't use for some people. The inconvenience cost adds up pretty quickly. Kinda makes a 5% increase in motherboard performance seem pretty irrelevant.

And Intel has a good history of actually fixing mistakes. FDIV bug -> replaced processors. VC820 SDRAM bug -> new board and free RDRAM RIMM.

Oh well. I'll stop ranting but hopefully you get my point.

Re:The most important thing...

[Eivind]

Only one thing matters. Warranty.

Maybe to Americans. In a lot of other civilized countries the consumer protection laws are good enough that warranties are completely irrelvant.

Let's take Norway as an example. You have trough law 2 years warranty, 5 years on capital goods that are meant to last significantly more than 2 years. Probably motherboards would fall under this category, but even if not, 2 years is decent for electronics.

If something happens in this period, excluding normal wear and consumer misuse, you have the rigth to a cost-free repair or a new replacement, at *your* option. (that's rigth, you don't have to accept the repiar but are free to demand a replacement part.) Furthermore if you do choose to accept the repair, you've got the rigth to cost-free borrow a similar item for the duration of the repair, if the repair is estimated to, or in actuality takes more than 1 week.

Warranties larger than this are not really needed.

Btw, there are exceptions to the rule that the customer can choose if he accepts repair or want the item replaced, namely when the costs are unacceptable, as in a large multiple of the repair-costs.

For example, if you buy a car, and after 5 months the clock stops working, you won't be able to demand the dealer replaces your "broken" car with a new one. In this case you'd have to accept the repair (with a borrowed-car for the duration if it takes more than a week, which it shouldn't in this case)

ECC Memory support is extremely important...

[callipygian-showsyst]

and that's why I got a D955XBK in my new dual-core P4 Extreme system. It's amazing how 99% of desktop computer users have NO IDEA if their memory flips a bit every now and then (let alone correcting it!). The Intel MOBO supports ECC memory.

Re:ECC Memory support is extremely important...

[jandrese]

Eh, unless your memory, power supply, or memory controller are marginal flipped bits are pretty damn rare. They do exist, I used to work for a place that had 50+ SGI Challenge XLs loaded to the gills with memory. We had to write special filters in the log watcher to filter out "Single bit parity error" messages, because there would be several of them per day. However, this was over thousands of memory sticks. On the one or two sticks you have in your PC, it's almost unheard of. You're far more likely to lose data due to an OS glitch. It's just extra money spent that probably won't correct a single bit in it's entire lifetime.

Anecdotal evidence: I have an old PII-400 at home with ECC memory. Thus far it has yet to correct a single bit error, despite being on 24/7 since January of 1998.

In other words, are you going to allow for a tiny amount of uncertainty (which may just be noise compared to the uncertainty you get from other parts of the system), for extra performance and cost savings? That's what you need to ask yourself. OTOH, parity checks are an easy way to spot a failing memory module, which can make debugging a hardware problem so much simpler.

Re:ECC Memory support is extremely important...

[ScrewMaster]

Now that's just not true. I know for a fact that Windows users are informed when their system "flips a bit now and then." The famed BSOD makes an appearance.

That Hexus site...

[callipygian-showsyst]

That Hexus site is one of the most anoying I've ever seen. Every other word is highlighted with a sponsored link that's often not related at all to the subject at hand. It seems to me that the whole point of that site is to have mildly-useful content simply as click fodder.

Still the old standards...

[Cyno01]

Abit, gigabyte and intel? Sorry, theres a new highend standard, especially for overclockers. If you havn't heard of DFI, they have some kickass boards, loads of features, and rock solid. I guess they have a dozen s775 mobos, and im sure they'll have i955 chipset boards soon.

OpenHardware anyone?

[SolitaryMan]

Looks like the time has come for OpenHardware and OpenMusic...

What I got out of it...

[ender-]

So basically what I got out of that article is that the performance of all three boards using default settings were almost exactly the same, and that unless I'm using Microsoft Movie Maker, a 2.4Ghz Athlon64 kicks the crap out of the 3.6Ghz P4 Systems with this new chipset.

Yup, I for one am glad that I went with an Athlon 64 over a P4.

Generally Satisfied with ABIT Quality

[dsci]

While I don't have experience with these particular boards, I've have had very good luck with two ABIT models:

KT7E and KX7-333R

The KT7E has been running virtually non-stop for 4 or 5 years; the KX7 likewise about three years. Both have been used constantly for their service life running floating point, number crunching code.

Both boards have withstood cpu fan failures and powersupply failures; replacement of the faulty components and they're up-n-runnin'.

I'd recommend ABIT to anybody with the caveat that I did have some problems with the KX7 until a bios update was done. Since then..no problem.

Though I have to admit, my newer systems are ASUS A8N-SLI Deluxes with Athlon 64 X2's. ;)

those were very likely capacitors

[YesIAmAScript]

Not transistors.

And honestly, many mobo companies had problems with this at that time. There was a company making knockoff capacitors that appears to be high quality components. Many good mobo makers got taken by this. It was covered on slashdot.

I personally won't buy Abit again if I can avoid it, for different reasons. My high end from them is still working, but I had significant problems in the beginning and minor problem all along, and they just never even responded to my queries on their forums.

Not using SLI, but have ASUS boards

[dsci]

if someone can find another reason to buy a motherboard with an extra high speed PCI-e slot... that's another story though.

I did.

I'm running ASUS A8N-SLI's with Athon 64 X2's on board in a small cluster. I'm not using the SLI feature.

I got these boards because:

1. ASUS performance is solid with the dual core Athlon 64.
2. Dual GbE on-board; no purchase of second NIC to use as headless cluster.

I can tie one NIC into my existing LAN and keep the other for the dedicated cluster interconnect. I don't have a KVM, but I can ssh directly into any node if need be.

3. *IF* I can full around and offload SOME number crunching to the GPU(s), maybe the SLI will come in handy in the future? It's a thought I'll keep in the back of my mind.

Their support blows.

[Some Random Username]

My KT7 stopped working the other day. As soon as you boot the alarm screams at you non-stop, even though nothing is wrong in the health section of the BIOS. And it locks up after about 30 seconds of being on, even when you are just sitting in the BIOS watching the seconds tick by.

Their support of course says "check the CPU temp" and calls it a day. Despite me clearly telling them everything is fine in the PC health section of the BIOS, and that I had swapped out RAM, CPU and video card to make certain it was the motherboard. The fact that they can't even bother to read support requests, nevermind give useful support is ridiculous.

Not really.

[Some Random Username]

First of all, its not like a cryptographic hash at all, its like a checksum, because it is a checksum.

Second, do you care a pointer that used to point to one memory address suddenly and mysteriously points to another, leading to a segfault, or blue screen if its kernel memory?

Errr...

[HaydnH]

From the article: "Having to replace a motherboard usually necessitates a long-winded process of reinstalling an OS, programs, and personal data."

Since when has swapping a motherboard forced me to format my hard disks???

You've gotta be kidding. Aren't you?

[Just Some Guy]

FDIV bug

You mean the one they initially didn't want to fix unless you could somehow prove that it might affect you? I'd be hard pressed to think of a worse example of this magical Intel warranty support you're so proud of.

They may indeed be better now; if so, then you should have picked a case that didn't make people hate them.