SPCR

can the HDMI handle HDCP properly? What about TrueDTS and bitstreaming other hi-def audio? What about 3D?

Basically, version of HDMI is it?

too bad it was tested with 1600 memory @ 1333 mhz... it support 1866 mhz natively, and performs better with faster memory

this is one of the first matx boards with 64 gb memory support (when 16 gb sticks will be available)

that mean we we will see mini-itx fm1 boards with 32 gb memory support (but 16 gb sticks doesn't exists yet)

It's a good enough board for most, but I find their lack of UEFI and modern fan speed regulation really disappointing. I've been on Gigabyte for many years now but next up seems to be Asus.

"The new APU is also capable of Dual Graphics, essentially teaming up with a discrete HD 6450, 6570, or 6670 card in CrossFire."

If two HD 6450 cards are used (not in CrossFire mode) in the two PCIe x16 slots, how many monitors can be powered under Windows 7, assuming all monitors have DisplayPort/DVI/VGA inputs?

Six? Eight (if two onboard video outputs are usable in addition to the two discrete cards' six outputs)?

Thanks.

I'm surprised you haven't commented on the apparently dramatically inefficient power circuitry on this board. In idle, 8.1W is drawn through the AUX12V connector while the APU itself should only consume roughly 2W. At full load, this 100W TDP chip draws nearly 150W of power through the AUX12V connector, which means that even if all CPU power would be routed through this connector the motherboard VRMs have an efficiency of easily less than 65% at full load - i.e. at least 50W of wasted power.

This means Gigabyte really does need that 2oz copper marketing bull to dissipate the heat generated by the VRMs, because that tiny heatsink itself cannot remove nearly that much heat. I bet you that if you would measure the MOSFET temperature on this board, they'll be sizzling hot*. More heat also means more need for cooling and a higher electricity bill, things we don't necessarily like here at SPCR.

These are observations that are crucial to assessing whether Llano power efficiency is 'good' at all. There is no doubt that Intel has the more efficient (i.e. performance per watt) line-up today in an absolute sense, but the difference with Llano shouldn't be this big. That's just poor power circuit engineering from Gigabyte.

* Only with great self-control I managed not to make a remark about my stunning physique here

(heh, I posted this just after release of the article but my post took some time to be 'approved')

I missed that too. A 100W TDP and yet it is drawing 150W. Is this a mistake?

Those 150 watt have to be an error, as other reviews around the Internet get 150 - 160 watt for the whole system.

zomg wrote:too bad it was tested with 1600 memory @ 1333 mhz... it support 1866 mhz natively, and performs better with faster memory

None of our tests would've been affected by memory speed.

hedly wrote:can the HDMI handle HDCP properly? What about TrueDTS and bitstreaming other hi-def audio? What about 3D?

Basically, version of HDMI is it?

HDMI 1.4. So yes, yes, and yes.

max789 wrote:If two HD 6450 cards are used (not in CrossFire mode) in the two PCIe x16 slots, how many monitors can be powered under Windows 7, assuming all monitors have DisplayPort/DVI/VGA inputs?

Six? Eight (if two onboard video outputs are usable in addition to the two discrete cards' six outputs)?

I believe the IGP disables if you use more than one discrete card.

HW_mee wrote:Those 150 watt have to be an error, as other reviews around the Internet get 150 - 160 watt for the whole system.

Not a mistake. We tested it twice, once on our CPU test platform (see: http://www.silentpcreview.com/article1211-page5.html) and once on our motherboard test platform. On CPU load the difference was only ~10W (due to component differences).

Note that we used CPUBurn and FurMark, synthetic stress testers; doing something real, encoding video with HandBrake, used 40W less and was on par with a Core i5-750. I don't think anyone who's thinking of buying an FM1 chip would do anything more demanding on the APU than that...

It might not be a mistake in itself (i.e. the measurements may be correct), but don't you agree that 150W power draw for a 100W TDP chip - TDP that is defined by that company as the worst-case continuous power draw under any load - would indicate massive losses along the line?

Lawrence Lee wrote:

HW_mee wrote:Those 150 watt have to be an error, as other reviews around the Internet get 150 - 160 watt for the whole system.

Not a mistake. We tested it twice, once on our CPU test platform (see: http://www.silentpcreview.com/article1211-page5.html) and once on our motherboard test platform. On CPU load the difference was only ~10W (due to component differences).

Note that we used CPUBurn and FurMark, synthetic stress testers; doing something real, encoding video with HandBrake, used 40W less and was on par with a Core i5-750. I don't think anyone who's thinking of buying an FM1 chip would do anything more demanding on the APU than that...

The results match what is expected from the earlier performed test, but still seem off.
SPCR seems to have the highest power consumption results in any Llano/FM1 review, but nice to know that you had doubts on the results as well.

Excellent review. Is the USB3 finally bootable? My 880/890GX aren't ! With CPU + GPU combined in the same die are huge coolers still necessary? I'm annoyed with Gigabyte's lack of fan control but I'm sticking with them because of Asus' overall quality problems.

And Gigabyte never seems to have ECC memory support.

Does Llano even have ECC memory support? I can't find anything on it

Nope. No ECC support in Llano.

Rebellious wrote:Excellent review. Is the USB3 finally bootable?

Yes, it's finally bootable, I'm booting XP \ 2003 from a USB 3.0 UFD connected to the USB 3.0 port.
(I wrote a small utility called IntegrateDrv that allows me to integrate the AMD USB 3.0 Host Controller and Hub to Windows XP \ 2003 Setup so I can install them directly to a USB 3.0 drive)

Tal Aloni