Shuttle spanks 8KHA+, well not really!

[spirits247]

I hate reading reviews that can spread misinformation, and this is one of them at Hothardware.

Epox and shuttle KT266A


Quotes:

<<"The 8KHA+ only offers a total of 3 fan headers, 2 of which are promptly used for the CPU and Chipset cooling, leaving room for one chassis fan.">>

<<"With that said, it would've been nice to see a minimum of 4 fan headers on the board to allow for easy integration of an additional chassis fan.">>


Erm, last time I checked I had 2 chassis fans, with a total of 4 fan headers on the board. Epox Tech won't be happy.


<<"When it came to overclocking these two motherboards, the difference was night and day. With our Thunderbird 1.2GHz. processor in the Epox 8KHA+, we were able to obtain a maximum overclock of 1440MHz. This was achieved by setting the CPU multiplier to 10.5x and the system bus speed to 137MHz. The increase of 240MHz. yielded an effective gain of 20%-not bad at all.

Next, we dropped the same processor and RAM into the Shuttle AK31 Ver.3.1 and began overclocking. We found the sweet spot for this board with the CPU multiplier set at 11.5 and the bus speed set at 136MHz. yielding at total CPU speed of 1564MHz. With an increase of 364MHz, we achieved an exceptional gain of over 30.33%!

Clearly, when it comes to overclocking, both of these boards are quite capable of getting the job done. However, in this comparison the Shuttle AK31 Ver.3.1 demonstrated a clear advantage in this round of testing. ">>


What? What? What? I have a friend with the shuttle and I was able to overclock just as far as he, but reached higher FSB. The CPU overclocking speeds were identical, like the majority of reviews have found. A case of PSU holding one back? Yes, I think so! The 8KHA+ probably requires a bit more from the PSU than the Shuttle at equal overclocks.

Every review I've seen puts the Epox down as a king of overclocking, but they never mention the reboot issue . Anyway, they seem to mention that this is a weird result:


<<"We feel it is important to note though, from what has been seen on the net in regards to the Epox 8KHA+, the overclocking performance of this particular board is not the norm.">>


Thanks!


At least in real time benchmarks and sandra memory the 8KHA+ came out on top every time, EVEN WITH THE 10% SLOWER CPU OVERCLOCK! HAIL TO THE KING BABY!

Sorry everyone, I'm feeling like a Pro-8KHA+ gusher today.

Sort of makes up for my last critical thread.

[dimmreaper]

It doesn't really matter, come January the Intel Northwood CPU and SiS 645 chipset combinations will be kicking all the SocketA arse there is, well until we see both .13 Athlons and KT333 . . . .

At least that's my oppinion . . . . and I'm not exactly a pro-Intel kinda of guy most days of the week . . . . though I'm not an anti-Intel kinda guy much either . . . .

[Aedan]

A case of PSU holding one back? Yes, I think so! The 8KHA+ probably requires a bit more from the PSU than the Shuttle at equal overclocks.

That suggests that the 8KHA+ is more sensitive and less efficent with it's voltage regulators. That would be a negative point but for the fact that only one board of each type was tested!

When it comes to overclocking component tolerances make a massive difference. Just because one board does not overclock well does not mean that the next board won't.

In their "Layout and Quality" section, no mention is made of Quality. They also refer to the forced cooling on the northbridge as "active cooling", which plainly it isn't. Can you see a TEC or similar heat pump on it?

Then, in their comparisons they use benchmarks. In the business benchmark, the difference is ~2%. In the content creation it's ~5%. 2% is within experiemental error. 5% might be outside, but without knowing it's hard to say. Again these look like they're single passes.

Basically, any conclusions reached based on any of the benchmarks are completely meaningless as we don't know the testing methodology used, nor the level of error involved.

Don't you just love pulling these "reviews" to pieces. It's actually very hard to do a sound review of a motherboard, and takes time to design proper tests and controls. I think StorageReview is one of the few places that starts to approach the matter in the correct way.

AidanII
P.S. Don't mention HardOCP, as they're even worse.

[KAIN]

IMHO I thought a fan could be classed as active cooling. If you just stuck a heatsink on something than that is passive cooling, but a fan draws air off of/onto something thus being active.

[Aedan]

Originally posted by KAIN
IMHO I thought a fan could be classed as active cooling. If you just stuck a heatsink on something than that is passive cooling, but a fan draws air off of/onto something thus being active.

A fan is classed as forced cooling, as it's forcing air across the device. Passive cooling is also known as convection cooling. Anything utilizing a heatpump is known as active cooling.

AidanII

[dimmreaper]

The points Aidan brings up are good ones, they just happen to be the reason why I'm not big on "reviews". I'd much rather ask several folks who own the product what they think about it. Not about specific numbers and benchmarks and the like, but rather if they are happy it or not and why . . . . but then I'm not a hard guy to please, all I ask is that the stuff does what it is advertised to do, and that it does so with out too much hassel . . . .

[dimmreaper]

Originally posted by AidanII
A fan is classed as forced cooling, as it's forcing air across the device. Passive cooling is also known as convection cooling. Anything utilizing a heatpump is known as active cooling.

To back Aidan, yup he is correct on this . . . .

Some reviewer knuckle head somewere along the line called a forced air cooler on a northbridge "active cooling" and every other knuckle head on the net has been using the same term incorrectly ever since . . . .

[KAIN]

Originally posted by AidanII


A fan is classed as forced cooling, as it's forcing air across the device. Passive cooling is also known as convection cooling. Anything utilizing a heatpump is known as active cooling.

AidanII

I stand corrected and apologise for my ignorance

Thanks for the info though...

[EPoX Tech]

I have read the article at Hot Hardware just the other day. It made for quite an interesting read. I know which mainboard I'd rather have at the end of the day

In response to Aidan II's comment about the HotHardware review stating power regulation being not as efficient on 8KHA+, this is incorrect. Nice point AidanII.

Power regulation is more sophisticated on 8KHA+ with a three phase arrangement which whilst more hungry for power does lead to a far more stable supply of phase regulation over any two phase solution.

Bring into play EPoX's other designs based on quad phase solutions and the theory is further qualified.

I won't be drawn into any debates on board manufacturer versus board manufacturer at the end of the day

[Aedan]

Originally posted by KAIN
I stand corrected and apologise for my ignorance

Nah, you're not ignorant. It's the people who push the term without really knowing what it means who are to blame.

AidanII

[spirits247]

You know what guys; reviews of motherboards are A LOAD OF CRAP!

Each one seems to contradict what the last has said, each one comes to different conclusions, each one judges a motherboard within an hour when we know it takes more like 2 weeks.

If I didn't know any better, I'd say some of them were paid to give bias opinions.

I know it's natural to defend the motherboard you have purchased, so I'm going to act natural!

The latest review over at Toms Hardware goes on to promote the Epox 8KHA+ saying:

The board is the leader of the pack in performance, and takes the gold in several benchmarks.

Page containing above quote

Indeed, after looking at the benchmarks the 8KHA+ came first/joint-first or second/joint-second in 12 out of 17 tests!

In the others it's like 3rd and/or 0.1-1% different (meaningless) from the top - fantastic

The ONLY test it did badly in was compiling Linux:

Linux Test

Something weird is going on there, I mean 30+ seconds slower than the Asus KT266A and nearly the same as a KT266???? Erm, No.


Overall some of the tests are a joke and quite meaningless. Some of the boards are pulling in identical results or within 0.1-1% yet being placed right down the list???? What a load of crap, only a 2% or higher difference is worth mentioning, let alone plotting!


Just like AidanII said, the difference is so marginal that the majority of ordered benches don't really mean anything at all, but for the sake of the review it seems to mean a lot!

Anyway, after giving the 8KHA+ a good intro with phrases like it's the " rage with particularly ambitious AMD users" they then rip it to pieces in the conclusion!

The Soltek is an ideal pick for any hardcore overclocker. Even at higher clock speeds (overclocking) and aggressive memory timing, the board runs much more stably than, say, the much-hyped Epox EP-8KHA+.

Page containing above quote

They make the above statement without any evidence - nowhere in the review do they mention testing stability at default or overclocked speeds. What's worse is that the above statement reads as though the Soltek is more stable at overclocked speeds than the Epox is at default!

Blasphemy! Another review said the Epox was one of the few boards that passed their 100 hour stability test, yet others make special mention of it’s stability and from using the board myself it certainly does not lack stability and that’s with Win ME!

Even at high-overclocked speeds, I may have the reboot issue but stability is still 100% in my tests.

To be fair, I think most the KT266A boards seem really stable – the chipset rocks.

The lesson to be learned is don’t trust any one review. Read them all, have a read around the forums and you should get a good feel about what’s right for you

[spirits247]

In relation to the above, this is the ONLY benchmark the 8KHA+ did badly in (large meaningful gap between results):

Linux Compiling test

Some 30+ seconds behind the Asus!

Yet it performs great in other tests - anyone explain this?

[spirits247]

Here are some more pointless benchmarks:

Soyo Dragon + at viahardware review

Soyo Dragon + vs Epox at gotapex

See what I mean? Four different reviews in this thread, four different set of benchmarks and different order/gaps between the machines.

The above 2 make the Dragon + look like different boards. One minute it's behind the others with an XP, and the next it's storming in the lead!

And just look at our Epox. In the Toms Hardware review it's coming out on top in the majority of tests over the dragon, in the gotapex it just plain gets slaughtered!

I'm through with reviews, most of them seem either bias, unfair or contradictory. I've read about 6 now and they all have different benchmarks, conclusions and results - what's the point?

I’m just going by price, features, forums and customer service from now on.