No FireWire

Dear MikeC,

as a µATX board, I mean it cannot (and it shouldn't) be compared with the full ATX M4A78T-E, but with a more similar one, like the M4A78-HTPC/RC with reference to its power figures are indeed more coherent: better yet, why (in the silence perspective that belongs to SPCR) haven't you checked the DFI with low power 4850e/5050e, or even 705e processors, as in the M4A78-HTPC/RC's article?

Regards,
Luca

some nice little touches but plenty of deal breakers for me- no esata, crappy layout, high price and bad efficiency.

quest_for_silence:
matx and atx- i can't see how they're any different if they use the same chips, my guess is that in this case the power circuitry has the biggest effect on efficiency... maybe.

Also, the board is meant mainly for overclockers and gamers, so it's relevant to test with high power components. It's easy to extrapolate what would happen with lower power components -- the differences in power between the DFI and the Asus would get smaller.

quest_for_silence: also, this is an AM3 only MB. You can't put an 4850e in it as that's an AM2 CPU.

Thanks for the review. Being the only AM3 mATX board out there, I've been looking at it. But the price and dual PCIe x16 slots turned me off. Although the BIOS undervolting options look nice. I don't suppose you could run some undervolting power tests?

Also, if the hot VRMs are possibly causing the higher power draw, what if you used a top-down heat sink or for a test, just positioned a fan to blow over them to see if that helped bring down the higher power draw.

Thanks!

I know, I know: I just wasn't sufficiently clear (english is not my native language).

This M3H5 is just an AM3 re-issue of the slightly former M2RS, so a test may be done with this one, or as suggested using the 705e/905e.

Regards,
Luca

The difference lays in where the chips are, and how much high is their temperatures: statistically power efficiency greatly decrease with high temperature, as maybe pointed out even by MikeC.

So, with reference to the full ATX Asus, the DFI is less efficient: but is it significant? Instead what about the more similar µATX ASUS? If I need a µATX board I would like to know if it is up to its pars.

Regards,
Luca

For sure, but any comparison should have been made with reference preferably to other µATX boards (see the 720BE power consumption on the M4A78-HTPC/RC: it's around 20w more on cpu load, nearby the difference between the AM2+ M3A78-T and AM3 M4A78-T).

Anyway, if you would ever test another AM3-HD3300 µATX board I'd have clearer, more definitive findings.

Regards,
Luca

Because "silence", "high performance" and silent are not mutually exclusive. There are a lot of ppl out there that need high performance PCs and still wanting them as quiet, efficient and silent as any other silent PC enthusiast here.

a poor argument, let me ask you this: who designed the dfi board?
asus has used the same chips and have produced a board with significantly better efficiency, sure the matx form factor is more limiting, but the layout of the chips isn't that different between them and i'm sure the reviewer pointed out the odd nb positioning- dfi's choice.

you say higher temps reduce efficiency, but the opposite is also true- bad efficiency leads to higher temps. in the end, it's the designer/engineer/maker who is responsible for the performance of their motherboard.

edit:
i think i might have something.
the larger atx board has more space, allowing it to support more sophisticated power circuitry. asus has a thing called 8+1 phase power, and judging by the pics, the dfi has a 4+1 thing going on, perhaps this makes a difference?
and maybe the larger surface area in general just lets thing breath more easily... i don't know

I really think no eSATA or firewire should be listed in the cons section. This leaves you with no direct high speed access to external storage, which is to me - and I guess many enthusiasts - a rathe big con. Which should be readily apparent especially if you only check out the pro/con section.

done.

But consider that eSATA is simply SATA with a different connector; a simple adapter SATA cable w/ captive PCI cover port is all that's needed for eSATA. I agree, tho, that not having it integrated on the back panel is a bit of a surprise for an enthusiast board.

As for firewire, it is starting to disappear from a lot of boards.

This is a matx dual-VGA board aimed at the performance gaming community. Many people won't have any PCI slot covers to spare. In that case, the only option would be an eSata connector on the front panel of the case, I guess.

Aren't "silence and silent" nearby the same thing? Maybe a typo?

Anyway, I never said so, I made a somehow different consideration: if you cope the above quoted requirements with "small form factor" (i.e. µATX or smaller footprint), they become more and more mutually exclusive the more the dimensions shrink, as laws of physics can't be overcome.

Regards,
Luca

It's not an argument, it's a matter of facts.

Let's see the DFI LanParty JR 790GX-M3H5's layout:



versus Asus M4A78-HTPC/RC's layout:



You see that without the need for the second graphic card slot the ASUS'es layout seem to be more rational (and therefore more effective?).

Well, let's see some relevant power figures:



Comparing µATX boards each other, and then the ATX ones each other, we may argue that the µATX form may be a major limiting factor, and it should be investigated (at least to definitely say if the DFI has a relative poor or not so good power efficiency, with reference to quiet computing).


You know that efficiency is declared for a given temperature range, so even such an hypotethical poor efficiency is a given percentage for a given temperature, and if you increase the latter, things are simply going worse. In fact, as you eventually note:


Yes, size matters: heat dissipation is in direct ratio to physical dimensions, isn't it? And so, for some respects, the power efficiency.

Regards,
Luca

that's correct, my mistake.
what i meant was that dfi designed the board, so if it isn't as efficient then it's their problem.



the temps show that the nb temps are about the same, the dfi is about 10C higher in sb temps, and a massive 17C higher in vrm temps.
now the similar nb temps suggest that the positioning hasn't increased its temperature, since its the same as the asus board even though it has a heatpipe going directly into a hotter heatsink. god knows what's with the sb temps, the heatsinks seems similar sized and on both boards they're far away from any other heat generating sources (no video cards were used in these tests)- perhaps some bad thermal goop. the vrm temps however, can be explained by the different heatsinks used, and the design of the power circuitry- both things are directly chosen/designed/engineered by dfi.



for heatsinks yes, but for motherboards i'm still not that convinced. i believe that if makers wanted to and tried to, they could make a matx board just as efficient as an atx board (i can see asus shoehorning their 8+1 phase thingy into a matx board... but then again i'm no engineer).


the great thing about matx is that they tend to be cheaper, but this sometimes leads to cheaper components and no vrm cooling
the dfi is more expensive than the asus, and is less efficient, the smaller size simply isn't a good enough excuse for me.

Hmmm - I'd love to build a very quiet µATX gaming rig out of that - two ATI cards with Accelero S1s on them, then a duct over them with a 120mm fan blowing through the duct towards the rear end of the case, all that combined with a big Phenom II in an Antec NSK3480 - sweet. Of course it could be difficult to route all theses cables, but we love challenges, do we?

Looks like you could get a nice performance-per-m³-figure with this board...

Now I would just have to convince my significant other that I needed a new toy - that would be the hardest part

For sure, but to be fair according to me it has to be compared to similar products.

It may worth to give a look to the more similar (as for form factor AND price) M4A78-HTPC/RC: SB temp is 53°C (-7°C), NB temp is 66°C (-20°C), VRM's temp seem to be identical (or nearby: even the VRM seem identical, a 4+1 if I'm correct).

With respect with the above quoted data, it should seem that the different positioning may affect NB's temp, even if the HD 3300 is more powerful than HD 3200.

And/or super imposed by the form factor, I guess.

Well, as said, the law of physics can't be overtaken by human will: the µATX form factor has more thermal limitations than the ATX (a smaller surface is able to dissipate less heat than a larger one, and this is a fact, a smaller surface may host less chip than a larger one, and so on).

The smaller size isn't an accident (or an excuse): it's a capital point of the whole project. This is my point: apples to apples, µATX to µATX.

If in case then you may discuss if the choice of Crossfire on a µATX motherboard is savvy or not.

But anyway, I have not to convince anyone of some evidences: if you (or MikeC) don't like that mobo, it won't be a problem.

Regards,
Luca

Maybe a positive pressure SST Sugo 03/04 could cool such a rig more effectively (but the toy will cost a lot more with the SG04H).

Regards,
Luca

the asus M4A78-HTPC/RC uses a different nb and sb.
and also, spcr reviews of am3 cpus on am2+ motherboards have given significantly higher power draws.
from the asus M4A78-HTPC/RC review:
As it turns out, the M4A78-HTPC/RC is among the lowest power AMD boards we've encountered, coming close to matching the Gigabyte's 740G-based board in idle power consumption. Power draw on load was also excellent
and
In our M4A78T-E review, we noticed a significant power consumption discrepancy of up to 22W when using an AM3 CPU on an AM2+ motherboard. Unfortunately the M4A78-HTPC/RC is also afflicted by this anomaly, so it doesn't appear to have been an isolated case

the hd3300 is used in both the dfi and atx asus.

i'm not arguing that the position of the nb doesn't affect it's cooling, but in the case of the dfi and atx asus- there isn't a large difference in temps.

matx is smaller, but the pcb itself only dissipates so much heat (read- hardly any). heatsinks can be made to overhang- if the manufacturer chooses to do so, there is plenty of space for larger heatsinks on any motherboard.

my point about the size is that it's the main thing it has over the asus, and frankly i don't believe the decrease in efficiency is a good enough trade-off.

edit- yes i have given up on quoting, its a pita!

Not quite. eSATA has a higher minimum transmit potential in order to handle longer cables and greater inference. Taking normal SATA outside the case may not work properly and greatly increases the chances of signal errors.

I'm a bit tired to repeat the same whole thing: prolly I have to do some course of english to communicate effectively on SPCR.

Look at the data, figures are for 20w more for the µATX ASUS 780G at load, with reference to µATX DFI: how much is the power leakage for an AM3 cpu on an AM2+ board? About 20w, as pointed out by the comparison of Asus M3A78-T and Asus M4A78T-E (just some messages above). How much is the power consumption difference at load between a 780G and a 790GX? According to some reviews (as Tom's HG) it measures under 4w. Do that expensive M4A78-HTPC/RC uses different VRM? No, they seem pretty the same. And so on. It seems a lot of coincidences to leave them out, don't it?


Pick up another µATX AM3 790GX mobo and measure it (and, incidentally, that's all I suggested to MikeC): if you'd find significant power savings, then you'd have found your small board of choice (naturally, if you don't mind about Crossfire).

Well, I give it up: have a nice day porkchop and, please, use less bold (not all people are presbyopic).

Regards,
Luca

i understand what you're getting at, factoring in everything the two matx boards do in theory have similar power draws.
i still believe that the design and component choice has a larger effect on efficiency than physical size, size will sometimes limit choices but there's always a way around it (think mini itx), remember that they are asking for a premium here.

sorry about the bold if you don't like it- i just thought it was easier to read this way.

i will have good day, and i thank you for debating this with me- it has made my day go by a little quicker which is always welcome on a work day

Well there aren't many 790GX AM3 µATX boards out there, it's this one and one from Jetway with both DDR2 and DDR3. The layout is almost identical, and it has a eSATA port. Newegg link, yes it's ugly.
Placing the NB close to the back is typically 790GX (not all though), Biostar looks like this too.

MikeC: The Jetway board would be perfect for comparing the power draw of the possible AM2, AM2+, AM3, DDR2 and DDR3 combinations since it removes a lot of variables when using the same motherboard. Sorry if I'm repeating myself, I may have posted this before.

Maybe this other Biostar could be more interesting (just for comparison purposes), since it does Crossfire and it is an AM3 one (as the DFI and Jetway), but it's amazing that its layout looks like more similar to the ASUS M4A78-HTPC/RC' one.

Regards,
Luca

P.S.: more than the "vexata quaestio" of an AM3 cpu in an AM2+ socket, I'd also like to find out any µATX mobo with 8+2 cpu's VRM, to eventually realize whether I'm a moron (as someone have to tell me that two mainboards are different, while I've already pointed out that they are different), or I'm just unable to properly understand english.

A mATX performance board aimed at gamers is still mostly a strange idea. They are also way overpriced. Who actually buys these then? A good midrange ATX board without the gimmicks will cost less, be much more versatile and perhaps more efficient and reliable.

Inexcusable.

From a silence perspective, yes, indeed.
Nonetheless the web is full of guys with gaming rigs equipped by crossfired HD4870s or GTXs in a µATX form factor (to say, something like Shuttles, or Sugos). Nice toys, tiny and angry, if you like the genre: if they was sport cars, maybe they'd be some Lotus.

Not so overpriced for what I said above (and even Jetway or Biostar come for $100-120 and more: not to mention ASUS). Anyway, they may be used even as HTPC giving some satisfaction.

Regards,
Luca

Well, aside from the differences and the space concerns already mentioned, looking at the accessoires picture this does not seem to be incuded. Which is not understandable at a price point of $140 (other vendors, e.g. Gigabyte, do include such a connector).

Dual PCIe x16 always costs more (even if it's dual x8 ), but I think S775 nVidia 9300's are just as overpriced, even though they cost less than this DFI.