Asus M3N78 Pro: Geforce 8300 & HybridPower Explored

Post by **MikeC** » Sun Jul 13, 2008 8:43 pm

Post by **CA_Steve** » Sun Jul 13, 2008 9:09 pm

Great review, Lawrence - disheartened to see the HybridPower results. I hope this can be improved down the road thru driver updates.

porkchop · Post by **porkchop** » Sun Jul 13, 2008 10:00 pm

i guess it can't be helped since this is an atx mb, but i'm very disappointed with the rear io ports- the m3n78-em(the matx version essentially) has an extra vga, dvi and esata port.

Scoop · Post by **Scoop** » Mon Jul 14, 2008 12:01 am

"does AMD even have any 140W processors in production?"

The new Phenom X4 9950BE has a TDP of 140W. Not that it draws that much but I guess the board at least claims to support it. But some may know that other IGP boards by many manufacturers for AMD CPUs have claimed support for processors that would actually literally blow the board apart.

Cistron · Post by **Cistron** » Mon Jul 14, 2008 12:20 am

Now we gotta wait for the mini-ITX versions.

Elvellon · Post by **Elvellon** » Mon Jul 14, 2008 1:01 am

Erm... Any chance of a GTX260 review now that you've got it? Or are SPCR VGA reviews self-limited with ~100W products?

Tzupy · Post by **Tzupy** » Mon Jul 14, 2008 4:23 am

Interesting review, thanks SPCR! I didn't know that the nVidia 8300 chipset is out.
But I'm dissapointed in the performance, I would have preferred it to use 10W more power, but have twice the 3D performance.
A good chipset cooler like the TR HR-05 IFX would have been capable to handle the extra 10W of heat.
Adding a gfx card to get decent performance adds a lot more power, while the Hybrid power isn't really working.
IMO it would have been more interesting to find out how much the IGP can be overclocked, not the CPU, a 4850e is good enough at stock.

Post by **MikeC** » Mon Jul 14, 2008 5:45 am

Elvellon wrote:Erm... Any change of a GTX260 review now that you've got it? Or are SPCR VGA reviews self-limited with ~100W products?

Yes, a review will come.

mikesm · Post by **mikesm** » Mon Jul 14, 2008 9:23 am

Mike, nice review, but it's missing a couple key points that a lot of HTPC users really want to know about this board:

1) This board is supposed to enable HD audio over HDMI so it can support 8 channel LPCM and possibly TrueHD, etc... when connected to a compatible reciever via HDMI. This is a huge benefit over the 780G which is limited to SPDIF quality audio over it's HDMI port. Please test this!

2) MPEG2 playback, particularly wrt to de-interlacing quality for 1080i material. You would think that if a card does VC1 and H.264 playback well, it would work great for old MPEG2 content, but is far from the case, usually because all HD-DVD and Bluray media are natively progressive, so de-interlacing is not tested. However ATSC video is commonly sent as 1080i, so the de-interlacing matters a lot. Please test that also...

Thanks again for a great review, and I really hop you'll consider these additional tests...

Elvellon · Post by **Elvellon** » Mon Jul 14, 2008 9:47 am

Anyway, thanks for a great review above anything else I see on the net. Seems like Phenoms are unneeded for an AMD video-decoding-limited PC with hardware acceleration.

hightower · Post by **hightower** » Mon Jul 14, 2008 12:23 pm

Could it be that GTX280 still have poor drivers for hybridpower - did you contact Nvidia?

Maybe uou should've tried it with a 9800GTX also...

SammyB · Post by **SammyB** » Mon Jul 14, 2008 1:58 pm

I got mine on July 9 but and have been fighting the dreaded 1 long beep, 3 short beeps problem. My AMD Athlon 64 X2 5000+ (DO) 512Kx2 65W SocketAM2 wasn't on the approved CPU list for this board so I had to buy a triple core CPU that was on the list but that didn't solve the problem.

Tried two different monitors with two different cables. All I get are those danged beeps.

smilingcrow · Post by **smilingcrow** » Mon Jul 14, 2008 4:45 pm

hightower wrote:Could it be that GTX280 still have poor drivers for hybridpower - did you contact Nvidia?
Maybe uou should've tried it with a 9800GTX also...

HybridPower works very well with the combination of the 780a SLI and 9800 GTX according to Legitreviews.

Monkeh16 · Post by **Monkeh16** » Mon Jul 14, 2008 6:17 pm

SammyB wrote:I got mine on July 9 but and have been fighting the dreaded 1 long beep, 3 short beeps problem. My AMD Athlon 64 X2 5000+ (DO) 512Kx2 65W SocketAM2 wasn't on the approved CPU list for this board so I had to buy a triple core CPU that was on the list but that didn't solve the problem.

Tried two different monitors with two different cables. All I get are those danged beeps.

Check your RAM.

Shemale · Post by **Shemale** » Mon Jul 14, 2008 7:30 pm

smilingcrow wrote:...HybridPower works very well with the combination of the 780a SLI and 9800 GTX according to Legitreviews.
[pic]

Something is fishy about the power consumption. From when a system with a 9800GTX running in 3D consume only 117W ?

.But in performance mode appear to be working, only if you don't move the mouse.

...For load measurements we ran World in Conflict at 1024x768 with medium game settings to see what the peak Wattage was during game play....

...A quick bump of the mouse and the system would change profile settings and the video card would power back on....

Spare Tire · Post by **Spare Tire** » Mon Jul 14, 2008 8:49 pm

I don't get it. You state that HybridPower only works with 9800 GTX, then you go on to test with GTX260 and say HybridPower doesn't save power. Well obviously.

Shemale · Post by **Shemale** » Mon Jul 14, 2008 9:22 pm

HybridPower work with GTX260, acording with nvidia:
http://www.nvidia.com/object/hybrid_sli_desktop.html

mattthemuppet · Post by **mattthemuppet** » Mon Jul 14, 2008 9:34 pm

great to see such a low power HTPC (or basic desktop) mobo, thanks for the review. The power consumption results are the complete opposite, but with a similar margin, to that found at the TechReport. They did use a Phenom, more memory and a raptor, but i would have though that would have just increased the no.s, not the difference between them. They also used a mATX board too, which is more comparable to the 780G board tested by both of you.

Any ideas?

nikolas · Post by **nikolas** » Tue Jul 15, 2008 12:18 am

Thank you for your nice review.

You have mentioned that at idle the system draws 37Wh, whereas by undervolting and underclocking the CPU to 1.5Ghz @1,1V the overall power draw was much lower.

However, according to your chart, the overall power consumption while undervolted and underclocked at idle was higher, (at 43Wh.)

Does that mean that the cool &quiet technology is so power efficient that at idle the system consumes less than a undervolted and underclocked CPU with cool and quiet disabled?
Or is there something wrong there?

SammyB · Post by **SammyB** » Tue Jul 15, 2008 1:26 am

Monkeh16, thanks for the advice. I have the Kingston DDR2 667 memory (2x 1GB sticks) and walked each stick separately through every slot cycling power each time. Same result, 1 long, 3 short beeps.

smilingcrow · Post by **smilingcrow** » Tue Jul 15, 2008 2:20 am

Shemale wrote:Something is fishy about the power consumption. From when a system with a 9800GTX running in 3D consume only 117W?

But thatâ€™s in Hybrid SLI Save Power Mode; it shows 194W for the 9800 GTX by itself.

renethx · Post by **renethx** » Tue Jul 15, 2008 3:52 am

M3N78 PRO supports Hybrid Power, but every graphics card supporting it is of length 10.5" that effectively blocks two SATA ports. I wonder what kind of morons designed this board.

amyhughes · Post by **amyhughes** » Tue Jul 15, 2008 7:00 am

nikolas wrote:Does that mean that the cool &quiet technology is so power efficient that at idle the system consumes less than a undervolted and underclocked CPU with cool and quiet disabled?

CnQ drops the CPU to 1 GHz. Underclocking disables CnQ, so the underclocked system is running faster at idle than the the stock system.

QuietOC · Post by **QuietOC** » Tue Jul 15, 2008 10:20 am

mattthemuppet wrote:They also used a mATX board too, which is more comparable to the 780G board tested by both of you.

Any ideas?

The chipsets are just one factor in motherboard power consumption. The VRM design is much more significant factor. Also voltages for the memory and other chips can make a significant difference.

amyhughes wrote:CnQ drops the CPU to 1 GHz. Underclocking disables CnQ, so the underclocked system is running faster at idle than the the stock system.

No reason why CnQ mechanism cannot be combined with (or even the method used for) underclocking. The mere reduction in clockspeed though only results in linear decrease in power at best.

It currently odd to be underclocking a multi-core processor. A higher clocked single core of the same design is more efficient. Intel could release a high speed 10W Core 2 Solo if they wanted. Unfortunately people want more and more of what they do not need.

smilingcrow · Post by **smilingcrow** » Tue Jul 15, 2008 10:34 am

QuietOC wrote:
mattthemuppet wrote:They also used a mATX board too, which is more comparable to the 780G board tested by both of you. Any ideas?
The chipsets are just one factor in motherboard power consumption. The VRM design is much more significant factor. Also voltages for the memory and other chips can make a significant difference.

Itâ€™s not unusual to see significant variations between motherboard power consumption figures for boards using the same chipset. There are so many ancillary components that impact consumption as well as the VRM etc as you suggest.

mattthemuppet · Post by **mattthemuppet** » Tue Jul 15, 2008 10:11 pm

smilingcrow wrote:
QuietOC wrote:
mattthemuppet wrote:They also used a mATX board too, which is more comparable to the 780G board tested by both of you. Any ideas?
The chipsets are just one factor in motherboard power consumption. The VRM design is much more significant factor. Also voltages for the memory and other chips can make a significant difference.
Itâ€™s not unusual to see significant variations between motherboard power consumption figures for boards using the same chipset. There are so many ancillary components that impact consumption as well as the VRM etc as you suggest.

Then I guess, when it comes down to this low level of power consumption, any conclusions must therefore be board, not chipset, specific. Given that the TechReport show a ~10W difference in 780G's favour and SPCR the opposite, that indicates that there can be up to 20W difference between boards based on the same chipset. In the system here, that's potentially an increase of 50% consumption at idle, which is incredible. Given that the processor, HDD and IGP should together consume less than this at idle, I find it really hard to understand how differences in VRM efficiencies and peripherals could account for such a relatively huge amount of power.

Post by **MikeC** » Wed Jul 16, 2008 6:57 am

mattthemuppet --

I'm not sure that you can compare the TechReport's results with ours like apples to apples.

1) The boards are different: Our Asus vs their Zotac
2) The system configs used are quite different: Our 45W TDP x2 4850e, 1GB RAM and 1W notebook drive vs their 95W TDP Phenom 9600, 2GB RAM and 8W Raptor HDD.

mattthemuppet · Post by **mattthemuppet** » Wed Jul 16, 2008 6:40 pm

MikeC wrote:mattthemuppet --

I'm not sure that you can compare the TechReport's results with ours like apples to apples.

1) The boards are different: Our Asus vs their Zotac
2) The system configs used are quite different: Our 45W TDP x2 4850e, 1GB RAM and 1W notebook drive vs their 95W TDP Phenom 9600, 2GB RAM and 8W Raptor HDD.

sure Mike, but that really just confirms what I said that any conclusion as to the power efficiency of particular chipset is only valid when applied to that particular board, which is a shame and a surprise. I really am surprised that a) a motherboard can consume so much power and b) different, basic boards (no fancy stuff) can differ so markedly.

As to point 2 - as the parts in each 2 way comparision remain the same for each board in that comparison, then the absolute difference should be purely due to the board, not the components. The difference relative to total power draw will obviously differ, but that's not what we're talking about.

mattthemuppet · Post by **mattthemuppet** » Wed Jul 16, 2008 9:33 pm

just to muddy the waters, saw this review on the fud - little difference at idle with 4850e if you ignore the OC results, bit different with a X4 9600, then all over the place at load.

Can't get my head around the differences between processors at load, guess that might be the VRM (or other) efficiency coming into play. Seems clear that, in that comparision at least

the 8200 is the more frugal at load.

hybrid2d4x4 · Post by **hybrid2d4x4** » Fri Jul 18, 2008 6:51 pm

Thanks for the review. I'm always eager to read up on how HybridPower and its AMD equivalent are progressing, and it's a shame about the results you guys got. I hope you'll follow up if nVidia explains why HP hasn't worked (or is this kind of outcome typical and legitreviews aren't so legit after all?)

Asus M3N78 Pro: Geforce 8300 & HybridPower Explored

Asus M3N78 Pro: Geforce 8300 & HybridPower Explored

Did you use the onboard VGA?

Re: Did you use the onboard VGA?

RAM Checked