1st PCIe slot (x16) only works at x2 with my Aorus 1080Ti...

[Tijgert]

I have this pretty Maximus X Hero (Wi-Fi AC) motherboard (Z370) paired with my Gigabyte Aorus GeForce GTX 1080Ti Xtreme Edition 11G.
The card is in the 1st PCIe slot and both parts support PCIe 3.0 speeds of x16.
The problem is that the card is stuck at x2 PCIe speeds and refuses to work at x16.

Even before hitting Windows and just checking in the bios, the bios says that the slot is running at x2 native, not x16.
The Windows install is fresh, latest drivers, but since the Bios already told me about this x2 limit I feel it's safe to assume it's not (Windows) software related.
In Windows in CPUz it says the same, running the PCIe slot at x2 only, but x16 capable.
Stressing the system with a game/benchmark doesn't up the PCIe speed, so no powersave mode afaics.
The motherboard ran bios 802, 1003 and 1101, all with the same problem.
I have not tried different videocard bioses (yet) as there are no updates available and trying non official bioses might kill something else besides my warranty.

I've reproduced this error/feature/bug on 2 of the same X Hero motherboards, so unless they're both identically broken it is not a hardware fault in the motherboard.
The videocard is fine, running at x16 in other motherboards and even x8 in the secondary PCIe slot on the Maximus X Hero, it's just stuck at x2 in the primary PCIe slot on this exact type of motherboard...

I don't like running the card in the secondary slot due to very restricted airflow, plus it's a brand new motherboard, it shouldn't be like this.

Can anybody shed some light on this?

There is one more bit of odd behavior:
I wanted to test the primary PCIe slot with some backup videocards and found they didn't give any picture at all.
An old GT220 and an Asus ENGTX560 both didn't give a picture on the monitor in the primary slot. But they worked in the secondary slot.
When I added both cards at the same time and checked the bios with the card in the secondary slot it said there was no card in the primary slot at all.
I didn't fidget with any of the PCIe settings, all vanilla.

 

[TheoneandonlyMrK]

Are you checking the bus speed via gpuz? Most motherboard and gpu combos drop slot speed in pure 2d mode ,gpuz shows mine doing this, or isn't this that?

 

[newtekie1]

What type of cooling are you using on the processor? You might have it too tight if it is a screw down cooler.

Are you checking the bus speed via gpuz? Most motherboard and gpu combos drop slot speed in pure 2d mode ,gpuz shows mine doing this, or isn't this that?

They never drop the number of PCI-E lanes. That is not a power saving feature and never has been.

 

[TheoneandonlyMrK]

What type of cooling are you using on the processor? You might have it too tight if it is a screw down cooler.



They never drop the number of PCI-E lanes. That is not a power saving feature and never has been.

sorry is it just the bandwidth then that drops?

 

[Yukikaze]

sorry is it just the bandwidth then that drops?

Yes, the PCIe link speed drops down to save power (often all the way down to Gen1 speeds), but the link width is never re-negotiated in runtime.

 

[erocker]

First, use the render test in GPU-Z to see if it goes to x16 under load:

If not, check for bent pins in the CPU socket (PCI-E pins are mostly located in the lower left to lower middle of socket) and/or bent connections in the PCI-E slot. Make sure you flash BOTH bios's on your motherboard to the latest (stable, listed on support page) bios.

The pinout chart for z370 starts on page 85: https://www.intel.com/content/dam/w...ts/200-series-chipset-pch-datasheet-vol-1.pdf

 

[mouacyk]

This can also be caused by having stuff blocking the electrical connections in the slot. Very dangerous... I killed a GTX 660 this way, the slot read at 1x in GPUz and crashes on high load.

 

[agent_x007]

Check pins inside socket.

 

[BadFrog]

There is one more bit of odd behavior:
I wanted to test the primary PCIe slot with some backup videocards and found they didn't give any picture at all.
An old GT220 and an Asus ENGTX560 both didn't give a picture on the monitor in the primary slot. But they worked in the secondary slot.
When I added both cards at the same time and checked the bios with the card in the secondary slot it said there was no card in the primary slot at all.
I didn't fidget with any of the PCIe settings, all vanilla.

New board? Return to retailer and get a new board. If past return policy, RMA

 

[coonbro]

did you install the cards driver from NVidia ? with no driver installed so to say safe mode that may be the fail safe limit it will do under default standard vga driver built in to Microsoft ? so with out a driver installed the card will not go in to its full power modes hence no x16

 

[eidairaman1]

Chipset drivers then gpu driver.

 

[jboydgolfer]

If you haven't yet, boot into the bios and manually set the PCI speed to revision 3. The band with was never something that was reduced as a power saving feature ,so that's kind of odd ,I wonder if there's an update for your bios

They released a bios update on 4/17/18

http://dlcdnet.asus.com/pub/ASUS/mb...I_AC/ROG-MAXIMUS-X-HERO-WIFI-AC-ASUS-1301.zip

 

[eidairaman1]

If you haven't yet, boot into the bios and manually set the PCI speed to revision 3. The band with was never something that was reduced as a power saving feature ,so that's kind of odd ,I wonder if there's an update for your bios

It is, my board drops to 1.1 idle and ramps to 2.0 Loaded.

I know a certain advanced setting in MSI afterburner wouldn't allow my cards bus rate to increase to 2.0 Max, it would stick to 1.1, as soon as I turned it off normal operation was restored.

 

[coonbro]

I'd also like to add put the card in its slot and pull it out 2 or 3 times to insure the ''golden fingers'' scratch in good contact in the slot . then if a long card check and see if something like the card is resting on something like a sata port or cable that may put pressure on the card that could make for improper seating of it in the slot .

brand new stuff may have a factory film on them golden fingers so when you have it out may wipe them down with a cloth and s spot of alcohol . maybe it will help ???

be sure all you psu plugs are pushed in firmly and correct . maybe getting a bad contact there as well

good luck

 

[erocker]

Thread is old, OP never came back. I'd assume he got it fixed.

 

[Tijgert]

So yeah, I’ve kept reading the replies but forgot to report what the problem was after is solved it.

Yes it is fixed. What was it? A glob of nail polish landed on one of the islands during delidding. Invisible at a glance I noticed a shiny island upon close inspection in bright sunlight. My trusty Aliexpress microscope showed it very clearly. Scraped it off, 16 lanes back again!

 

[jboydgolfer]

So yeah, I’ve kept reading the replies but forgot to report what the problem was after is solved it.

Yes it is fixed. What was it? A glob of nail polish landed on one of the islands during delidding. Invisible at a glance I noticed a shiny island upon close inspection in bright sunlight. My trusty Aliexpress microscope showed it very clearly. Scraped it off, 16 lanes back again!

Now that must be a fancy motherboard to have islands

glad you got it fixed though, even more so that you reported the results back

 

[eidairaman1]

Now that must be a fancy motherboard to have islands

glad you got it fixed though, even more so that you reported the results back

Unfortunate it landed on an ID pin

 

[Tijgert]

Now that must be a fancy motherboard to have islands
glad you got it fixed though, even more so that you reported the results back

Well, what else would you call the not-pins on the underside of the CPU?
It was a real bastard to find since it was clear polish and I looked straight at it and still could not see it. Only direct sunlight hinted at it because of the glimmer. Before that I exchanged the motherboard and even lined up a component test with likeminded enthousiats (which would’ve shown the CPU was the culprit).
I laid under the desk with pointy metal pins scratching the PCIe slot teeth/things.
I was all but convinced I killed the CPU while scratching off the silicon adhesive of the IHS...

Glad this all fizzed out with a whimper.

 

[rtwjunkie]

So yeah, I’ve kept reading the replies but forgot to report what the problem was after is solved it.

Yes it is fixed. What was it? A glob of nail polish landed on one of the islands during delidding. Invisible at a glance I noticed a shiny island upon close inspection in bright sunlight. My trusty Aliexpress microscope showed it very clearly. Scraped it off, 16 lanes back again!

Thank you for sharing your solution!

 

[jboydgolfer]

Well, what else would you call the not-pins on the underside of the CPU?
It was a real bastard to find since it was clear polish and I looked straight at it and still could not see it. Only direct sunlight hinted at it because of the glimmer. Before that I exchanged the motherboard and even lined up a component test with likeminded enthousiats (which would’ve shown the CPU was the culprit).
I laid under the desk with pointy metal pins scratching the PCIe slot teeth/things.
I was all but convinced I killed the CPU while scratching off the silicon adhesive of the IHS...

Glad this all fizzed out with a whimper.

i was joking,

but to answer your question , i suppose if your referring to the contacts/pads under the cpu, id call them contacts or pads? i dont know though, i dont think ive ever considered it to be honest