PCIe lane switching

[BlueBill]

It'll never happen, this project was dead from conception.

I'm curious what the OP was working on, but we'll probably never know.

[QinX]

If you actually succeed at making ASUS to write you a BIOS patch for a one off board, hats off ...

Unless you are golf mates with the CEO or something at a similar level, it is about as likely to happen as hell freezing over - it is a fairly significant engineering work you are most likely not ready to pay for and not economical for them unless you are willing to buy at least few hundreds/thousands of those boards.

I think you overestimate changing a BIOS, you see special BIOS for overclockers all the time, so why couldn't it be done for something like this, most manufacturers should already have the code, it's almost just a matter of implementing it for a specific board.

Now I'm not saying it is cheap and they will do it for free. But I also don't think they completly ignore it.
Unfortunatly you probably need to have some marketing related friends at said manufacturers to get inside. no way a simple email would do the trick.

It'll never happen, this project was dead from conception.

I'm curious what the OP was working on, but we'll probably never know.

Never say never, I believe that if the hardware is capable of it, then a software solution can be achieved, in addition to the extra PCB it would need.

I want to run 2 GPUs on a M-ITX motherboard for a product I am working on.

[BlueBill]

So why does it have to be Mini-ITX, a pair of video cards if powerful are going to require a fair amount of space and cooling. Seems a forum is looking for the same thing, but to no avail. It's simply too small a market and cooling would likely be an issue.
http://www.overclock.net/t/1515838/mini-dtx-hype-thread-itx-with-two-pcie-slots

It might help if you could explain what you plan to do with dual GPUs? There might be other solutions, would a dual GPU video card work like the ASUS ROG MARS 760?

[janoc]

I think you overestimate changing a BIOS, you see special BIOS for overclockers all the time, so why couldn't it be done for something like this, most manufacturers should already have the code, it's almost just a matter of implementing it for a specific board.

The big difference is that they produce and sell tens of thousands of those enthusiast boards. So the extra work required gets amortized in the quantity manufactured. I don't think anyone will be willing to do such mod for a one off, especially for a board that wasn't designed for it so it isn't a matter of flipping some flag somewhere. It is just too much hassle and risk. Writing the code, QA, support, you have mobilized a fairly large team of people and resources that have to be paid - all that because of a single board. The cost of that simply wouldn't be economical and it is a lot cheaper for the OP to find another solution for their need.

[BlueBill]

There's an ASRock H81M-DG4 which has two PCI slots.

[Billy Bones]

Hello everyone,

sorry to "necro" this old post, but I'm interested in the very same thing QinX is: two GPU PCIe cards into a single PCIe slot on a mini-ITX motherboard.

QinX, did you abandon this idea in the end?

Did you try out the SuperMicro RSC-R2UG-A2E16-A (http://www.supermicro.com/a_images/products/Accessories/RSC-R2UG-A2E16-A.jpg) or was it assuredly only compatible with some of SuperMicro's own server hardware?
Has anyone already emailed SuperMicro about that?

Is there any news about that BIOS code? Did you correspond with ASUS or someone else any further?

I found your build log on Overclock.net, QinX and it's very interesting. Nice work! You have quite the footsteps to follow.

Best regards,
Billy

[BurtyB]

It depends what your requirements are but I've successfully used a load of http://amfeltec.com/products/flexible-x1-pci-express-3-way-splitter/ to connect 3 PCIe graphics cards (with x16 to x1 adapters) to a mini-itx MB with a single x16 port which worked OK (for use with opencl with minimal communication going on).

[genBTC]

It depends what your requirements are but I've successfully used a load of http://amfeltec.com/products/flexible-x1-pci-express-3-way-splitter/ to connect 3 PCIe graphics cards (with x16 to x1 adapters) to a mini-itx MB with a single x16 port which worked OK (for use with opencl with minimal communication going on).

"for use with opencl with minimal communication going on" <- sure x1 is fine for this.
But for science, why not find out how fast she can go!

There is now an additional product from the same company: http://amfeltec.com/products/flexible-x4-pci-express-4-way-splitter-gpu-oriented/ Which uses an x4 host card branching out to a total of 4 physical x16 slot adapters (included). HOWEVER these seem to be attached by the same electrically x1 flat flex ribbon cables just like the other product. I was unable to find out which chip they were using; pic is not high res enough and its not itemized in the datasheet, but it is for sure a PLX chip.
(Large ,has a heatsink shown elsewhere, and datasheet/specs-page seems to indicate that each of the 4 are all addressable simultaneously and without a bottleneck).

It would be interesting to know how much these solutions cost, and if indeed it would be possible to design a better one, either with an x16 host card branching into 4 x4's, or the original intention of x16 into x8+x8. Either using a faster more expensive PLX chip to copy this solution, or with the quad ASMedia 1480 chips (how i found this thread). I have confidence in the BIOS bifurcation thing being possible.

I see no reason for this hardware build to be impossible, just a matter of why.
-Abei Villafane

[davidm71]

Hi,

I successfully modded my bios on an Msi x99a motherboard and 6850k on the last pcie x8 slot into two using a Supermicro 8x to 4x4x card. This lets me use my onboard M.2 drive at 4x and another nvme pcie drive on the last slot also at 4x! Only issue is sometimes randomly after a reboot the link speed of one of the connected devices drops to 2.5 gt/s from 8 gt/s. Looking for possible solution if there is one?

Thanks