Motherboard Mini-DTX petition thread!

[3lfk1ng]

Wouldn't 4 video cards be possible since there's two x16 slots?

If my math is correct, 8 GPUs.

 

[HeroXLazer]

Man imagine 2 single slot 1070, the one galax shows of last year (man it is 2017 ), water cooled inside a NCASE M1.. holy moly.

Or use a Titan XP and a GTX 1070 as a PhysX card

 

[Josh | NFC]

 

[jeshikat]

Or use a Titan XP and a GTX 1070 as a PhysX card

Is having a PhysX card still a thing?

 

[Kmpkt]

It is now that you can attach 8 video cards to a mITX board.

 

[HeroXLazer]

Is having a PhysX card still a thing?

Yep, the Geforce Garage video in the summer for deadmau5 had two 1080s and an extra one in his main rig out of five rigs.

 

[aquelito]

True, but those cases are still a small minority of the ITX case market. Plus like you said, it's very, very dependent on motherboard/BIOS/driver support.

Then we need a petition to generalize PCIe bifurcation !
It seems indeed standard practice for servers, why not for consumer products ??

At least you are sure to run both GPUs at 8x (instead of 16x and 4x).

Once again we can thank Asrock for being the only adventurous manufacturer out there.

 

[robbee]

Once again we can thank Asrock for being the only adventurous manufacturer out there.

Biostar has one too: http://www.biostar-usa.com/app/en-us/mb/introduction.php?S_ID=831#specification

It does seem to be quite low-end though, and I'm not too sure if I trust Biostar enough to buy low-end stuff from them.

 

[confusis]

ECS also produces low-end M-DTX

 

[Arboreal]

I've wanted a 2 slot board for ages, the Biostar H61 and H81 in that size seemed OK, but not high enough spec and not (from experience) a top tier manufacturer.
Bring them on, you have my vote
I'd be happy with a PCIe1x and 16x slot, but Z chipset please, decent audio, M.2 and on board wifi / wifi slot.
PCIe 16x and a better 2nd slot (4x or 8x in full length) would be perfect.
Lots of cases would accept these, and it would make my own plans even better.

 

[confusis]

Front page 'propaganda' here: https://smallformfactor.net/articles/editorial/opinion-m-dtx-ultimate-sff-motherboard

 

[CC Ricers]

I'd buy from Biostar if only they would improve in quality.

 

[Popstar]

What I really want is a FlexATX (229 x 191) to replace microATX. It would be a much better middle ground between ATX and mini-ITX.

 

[Necere]

What I really want is a FlexATX (229 x 191) to replace microATX. It would be a much better middle ground between ATX and mini-ITX.

Problem is, it wouldn't fit in most mini-ITX cases, while mini-DTX does. You could only use it in mATX cases, which defeats the point.

 

[Runamok81]

Why stop at 2?

 

[jeshikat]

But now we're getting close to microATX

 

[Popstar]

Problem is, it wouldn't fit in most mini-ITX cases, while mini-DTX does. You could only use it in mATX cases, which defeats the point.

If no-one was making mATX boards anymore, I assume people would also stop making mATX cases and instead make them for whatever the replacement was. It's not going to happen of course, but neither is mini-DTX.

 

[chx]

If you are talking to MB people could you please tell them that four SO-DIMM sockets instead of two DIMM would make a billion times more sense on an ITX board?

 

[Runamok81]

If you are talking to MB people could you please tell them that four SO-DIMM sockets instead of two DIMM would make a billion times more sense on an ITX board?

This please. I just don't see the need for full size DIMM anymore.

I do understand that higher quality RAM (better timings, ECC) is only offered in full size. But I think those benefits are overated? We're no longer in an age where RAM timings can net you a significant performance advantage. In all of the latest "RAM performance" articles I've read there is maybe ... 2% ? of performance to eek out between the bottom-barrel RAM and top.

SO-DIMM makes me happy.

 

[Ceros_X]

This please. I just don't see the need for full size DIMM anymore.

I do understand that higher quality RAM (better timings, ECC) is only offered in full size. But I think those benefits are overated? We're no longer in an age where RAM timings can net you a significant performance advantage. In all of the latest "RAM performance" articles I've read there is maybe ... 2% ? of performance to eek out between the bottom-barrel RAM and top.

SO-DIMM makes me happy.

I think the research recently has proven otherwise. At work, but this article summarizes findings on ram speeds yielding performance boosts: http://www.digitaltrends.com/computing/highspeed-ddr4-shown-to-improve-in-game-performance/ - they quote this article.