Oct 23, 2017 2:30:32 GMT -8

Reply Quote | Like

	eracoon Newbie	Right now I'm Offline I have made 11 posts I joined October 2017	My gender is Male	I currently have no messenger contact information filled out.	I currently have no social network information filled out.

Hi!
I have a “strange” problem to get my vga card running in another PCIe slot than the first one.
The system is very stable when I put the card into the first x16 PCIe port.

But when I try to use one of the other slots (2-5) my system gets very unstable (bluescreen), not recognize my GPU or Windows can’t load the driver.

This error is reproducible with every other GPU on slot 2-5.
Slot 1 is working fine every time with every card.
I tried everything to nail this problem down to the mainboard… trying multiple GPUs, using different RAM configuration (1 Module only too), installed multiple times Windows 10, everything on stock frequencies, multiple BIOS versions etc…
Everytime the exact same problem with the GPU in a different slot than the first one.

I have a single GPU machine, just want to put the VGA card into a lower slot to get a bit space between my Noctua NH14 TR4 and the VGA card.
Prevered of corse the second x16 slot (slot 4), I just tried the other one to see if the problem only exists on port 4.

I’m using a Samsung SM961 512GB SSD on the NVME port without problems, tried also all 3 NVME ports without any change to the VGA card problem.
So it looks like the problem is only on the PCIe slots with a VGA card and not on the pcie bus generally.

Do I have received a broken mainboard and need to open a warrenty case or is this maybe a general problem/bug with that board?
Or maybe some BIOS option I have to set?
I also tried to diasble IOMMU, same problem.

Thanks for your help!

My System:
TR 1950x
X399 Gaming 7
4x8 GB RAM (F4-3600C17Q-32GVK) (from the HCL)
Asus GTX 670 (GTX670-DC2T-2GD5)
NVME SSD Samsung SM961 512GB

Edited Oct 23, 2017 2:45:40 GMT -8 by eracoon

Oct 24, 2017 16:24:40 GMT -8

Reply Quote | Like

	echto Elite	Right now I'm Offline I have made 121 posts I joined September 2017 I've received 5 likes	My gender is Male	I currently have no messenger contact information filled out.	I currently have no social network information filled out.

There is a PCI version setting in the bios that defaults to auto. Try setting that to PCI version 2.

Oct 25, 2017 9:08:04 GMT -8

Reply Quote | Like

	eracoon Newbie	Right now I'm Offline I have made 11 posts I joined October 2017	My gender is Male	I currently have no messenger contact information filled out.	I currently have no social network information filled out.

I have tried auto vs. fix pcie 3.0, same problem.

I can try 2.0 too but my card has pcie 3.0, why should this be a problem.
Or lets say why has Gigaybte a problem with a 3.0 card in a 3.0 port?

Oct 25, 2017 17:42:51 GMT -8

Reply Quote | Like

	echto Elite	Right now I'm Offline I have made 121 posts I joined September 2017 I've received 5 likes	My gender is Male	I currently have no messenger contact information filled out.	I currently have no social network information filled out.

good question

Oct 26, 2017 5:38:06 GMT -8

Reply Quote | Like

	eracoon Newbie	Right now I'm Offline I have made 11 posts I joined October 2017	My gender is Male	I currently have no messenger contact information filled out.	I currently have no social network information filled out.

I got today an answer from the official Gigabyte eSupport with a very unsatisfying answer.

So that means that it is official not supported to put a single gpu into one of the lower PCIe slots?
Come on! That design concept is just very bad!

The new AMD platform has a very large socket and a high TDP.
Makes people using large coolers which get very close to the first PCIe Slot.

The fact that half of the RAM banks are on the left site behind the tower coolers makes this even more a problem.
So you get a heat point behind the cooler. The GPU which close the gap to the buttom and producing its own heat will definitely not help with that.

Especially the last RAM module which is directly in front of the heatcooler of the mainboard which is also getting very hot from the VRMs getting a lot warmer.

The both modules on the right site are at around 30°C and the one on the left behind the cooler is already at around 40°C and the last one in front of the VRM cooler hitting the 45°C without any OC.

So I already have a temp difference from around 10-15°C in my RAM module.
Moving the GPU reducing this to <5°C.
(Yes I already have a case fan directly behind the mainboard installed)

I'm not sure if I want to start getting into OC with that platform or switching over to another MB.
I was planning to get to 4Ghz+ with my TR 1950x, where I calculated with 100% more power consumption and heat from the CPU.
This + the higher VRM temps would maybe get my system instable because my last RAM modules and the VRMs getting to hot.

This is just broken by design… serious Gigabyte?
Or is the support answer just wrong?

Moving my CPU cooler 6mm up would not help with that problem... O_o

Although it’s not a standard setup to install only one PCIE VGA card on the second PCIE X16 slot or other PCIE slot, it should be OK to finish O.S. installation/boot up at such circumstance.

Customer can try to check with different VGA card if possible.

Please try to set Initial Display Output to the slot where the card is installed.

Please notice to customer that the cooler Noctua NH14 TR4 does has an offset mounting features that can move the cooler 0-6mm farther than the first PCIE slot:

noctua.at/en/nh-u14s-tr4-sp3

Customer can simply use such feature instead of having such trouble by using non-standard setup.

Edited Oct 26, 2017 5:41:35 GMT -8 by eracoon

Oct 26, 2017 17:00:50 GMT -8

Reply Quote | Like

	echto Elite	Right now I'm Offline I have made 121 posts I joined September 2017 I've received 5 likes	My gender is Male	I currently have no messenger contact information filled out.	I currently have no social network information filled out.

Oct 26, 2017 5:38:06 GMT -8 eracoon said:

I can understand your frustration, but please do not overlook alternate cooling solutions. You can always install a liquid cooler and mount the radiator in the top part of the case, or if its a small radiator, mount it on the side.

Edited Oct 26, 2017 17:01:40 GMT -8 by echto

Oct 27, 2017 0:33:09 GMT -8

Reply Quote | Like

	eracoon Newbie	Right now I'm Offline I have made 11 posts I joined October 2017	My gender is Male	I currently have no messenger contact information filled out.	I currently have no social network information filled out.

Oct 26, 2017 17:00:50 GMT -8 echto said:

Thanks for your respond, looks like you are the only one here in the forum who trying to help.

You are talking about AIO liquid cooler or custom ones?
The Noctua NH14 TR4 with 2 fans installed is more powerful than all the standard 360 AIOs that you can buy.
Only the first one with a bigger head spreader for TR4 from Enermax is a bit more powerful and even this is only about a very tiny difference in temperature.

So I don't see any reason why I should not go into OC with that cooler.
When I move my GPU down, I don't get any problems.

BTW it looks like setting the PCIE Gen fix to GEN2 helped so far.
My system is now stable for 2 days with that setting.

But not really a solution to slow down your components on a high end platform to get them stable with the stock configuration.
If there will be no fix soon this was definitely my last Gigabyte board I bought in my life!

Looks like the GEN2 setting is also affecting the M2 slots... so at the moment my NVME SSD gets around 1GB/s less read/write rates.

Edited Oct 27, 2017 0:38:17 GMT -8 by eracoon

Oct 28, 2017 18:13:54 GMT -8

Reply Quote | Like

	echto Elite	Right now I'm Offline I have made 121 posts I joined September 2017 I've received 5 likes	My gender is Male	I currently have no messenger contact information filled out.	I currently have no social network information filled out.

That is very odd and disappointing. Let's hope Gigabyte is aware of this and it's something that will be addressed in a future bios update. I have yet to go the M.2 route, but will soon if I see something good in a Black Friday sale.

Feb 3, 2019 1:49:01 GMT -8

Reply Quote | Like

Roger_Weihrauch
Guest

Hi eracoon

Sorry,but not sure if your problem is/had been already solved. I had the same issue: GPU only running in a lower port and additionally some freezes/blackouts on probable timings.
My gpu is a Gigabyte Aurous RX580 AMD,8GB.

I also tried any seting in BIOS, no success. What really helped in my case was the latest GPU BIOS update. This solved: usinf git in different pcie slots, no more blackscreens and also phoronix test suite is running noe in linux.

Hope this helps,
Roger

Feb 20, 2019 10:28:12 GMT -8

Reply Quote | Like

	eracoon Newbie	Right now I'm Offline I have made 11 posts I joined October 2017	My gender is Male	I currently have no messenger contact information filled out.	I currently have no social network information filled out.

Meanwhile my Gaming 7 Board died and ended up in a reboot loop... in RMA at the moment.
So I bought an Asrock Gaming 6 to get my computer back to work... of course the Asrock has no problem with my GPU in a lower PCIe Port.

Means that this is definitely a problem with that board and not a generell problem with the x399 AMD chipset.
Was definitely my last Gigabyte motherboard... I'm so frustrated!

Feb 25, 2019 4:01:04 GMT -8

Reply Quote | Like

	chickenripper Newbie	Right now I'm Offline I have made 2 posts I joined February 2019		I currently have no messenger contact information filled out.	I currently have no social network information filled out.

That reply is bullsh-t. I have 2 GTX 1080s in SLI in slot 2 and 4 of my Aruos Extreme X399 with a 2990wx.
I use the x16 slots for quad nvme cards.
Not the same mobo but I did the same on my older X399 Taichi without issues.

have you tried completely uninstalling the vid card driver before you move the card down?

ah just read that the board is cooked.. does it have a bios backup fixer thingy?

Feb 28, 2019 9:39:28 GMT -8

Reply Quote | Like

BrianUK
Guest

I use my GTX 1060 in the second x16 slot (slot 4 from the top), it works fine and I've never had any problems. I have 2 NVMe drives in the two long slots on the board

kshipper wrote: Apr 22, 2021 9:05:16 GMT -8
mcglynnj: You should make a post in the forums on the left here with as much info as you can about your problem(s) ..list parts used to build your machine. Did you build it yourself. What did you try to do to fix it etc.. *

mcglynnj wrote: Apr 22, 2021 0:32:48 GMT -8
Need help with a fault on my Z390 I AORUS PRO WIFI board and eSupport are just making life so difficult and unnerving for me... Been at this for almost a year...

rheady32 wrote: Apr 21, 2021 6:10:13 GMT -8
nvr buying gigabyte again. YOU ARE TRASH AND SO IS YOUR SOFTWARE. GO make case fans or something......... on second thought dont....