Author

Topic: GPU 1 on Gigabyte Motherboard is always generating HW errors (Read 137 times)

legendary
Activity: 1834
Merit: 1136
You can try this option, because only 5 GPUs are used. I have a ga970a-ds3p test kit, AMD X3-X4 CPU, 4GB RAM. On aliexpress, this kit costs no more than $ 60. I advise you to buy novice miners. The motherboard does not require BIOS settings.
member
Activity: 325
Merit: 42
Hello,

I have 5 RX580 8Gb cards under HiveOS.

GPU 1 is always getting HW errors... but the strange thing is that it's happening with any card listed on GPU 1 !



I first thought it came from the card or the riser but I figured out that this is not hardware related problem.

I thought it's the biosmod but I tried with the stock rom and it it does not change anything

I removed the GPU 1 card from the rig and replaced it twice with 2 different GPUs (different risers, power cable...) on the PCI slot. They all got a lot of HW errors…
When they are recognized as GPU 2/3/4/5 they are working without issues.

I also leaved the PCI slot of GPU 1 empty and only pu the cards on the other PCI slots of the motherboard: GPU 0 / GPU 2 / GPU 3 / GPU 4 / GPU 5



But I suppose the computer does assign the PCI slot of GPU 2 (PCIEX1_2) as GPU slot 1. Indeed there are always HW errors indicated in GPU 1…while physically there is nothing on this slot.

I configured the BIOS as indicated in the manufacturer's instructions. (https://www.gigabyte.com/MicroSite/462/mining.html)

I can't figure out what is causing this issue.. ? A BIOS setup? A hardware problem with the motherboard? It's a GA-H110-D3A (Gigabyte). Bios is up to date.

I am using HiveOs and Teamredminer (But also tried under other miner and SimpleminingOs)

Thank you 🙂


Looking at your picture of the mobo with risers it seems you have the first riser in the PCIe 16 slot and not in the PCIEX1_1 slot. The PCIe 16 slot should be empty.
In other words only use PCIEX1_1/2/3/4/5 slots.
Why PCIe 16 slot should be empty?
I had 1 such motherboard and video cards worked fine in various installations. And I also had 5 cards instead of 6, because for 6 cards I needed a more powerful power supply.

It does no harm to try it out. and for the PCIe 16x slot I probable would stick a videocard in directly.
legendary
Activity: 1708
Merit: 1615
Payment Gateway Allows Recurring Payments
Hello,

I have 5 RX580 8Gb cards under HiveOS.

GPU 1 is always getting HW errors... but the strange thing is that it's happening with any card listed on GPU 1 !



I first thought it came from the card or the riser but I figured out that this is not hardware related problem.

I thought it's the biosmod but I tried with the stock rom and it it does not change anything

I removed the GPU 1 card from the rig and replaced it twice with 2 different GPUs (different risers, power cable...) on the PCI slot. They all got a lot of HW errors…
When they are recognized as GPU 2/3/4/5 they are working without issues.

I also leaved the PCI slot of GPU 1 empty and only pu the cards on the other PCI slots of the motherboard: GPU 0 / GPU 2 / GPU 3 / GPU 4 / GPU 5



But I suppose the computer does assign the PCI slot of GPU 2 (PCIEX1_2) as GPU slot 1. Indeed there are always HW errors indicated in GPU 1…while physically there is nothing on this slot.

I configured the BIOS as indicated in the manufacturer's instructions. (https://www.gigabyte.com/MicroSite/462/mining.html)

I can't figure out what is causing this issue.. ? A BIOS setup? A hardware problem with the motherboard? It's a GA-H110-D3A (Gigabyte). Bios is up to date.

I am using HiveOs and Teamredminer (But also tried under other miner and SimpleminingOs)

Thank you 🙂


Looking at your picture of the mobo with risers it seems you have the first riser in the PCIe 16 slot and not in the PCIEX1_1 slot. The PCIe 16 slot should be empty.
In other words only use PCIEX1_1/2/3/4/5 slots.
Why PCIe 16 slot should be empty?
I had 1 such motherboard and video cards worked fine in various installations. And I also had 5 cards instead of 6, because for 6 cards I needed a more powerful power supply.
member
Activity: 208
Merit: 46
Hello,

I have 5 RX580 8Gb cards under HiveOS.

GPU 1 is always getting HW errors... but the strange thing is that it's happening with any card listed on GPU 1 !



I first thought it came from the card or the riser but I figured out that this is not hardware related problem.

I thought it's the biosmod but I tried with the stock rom and it it does not change anything

I removed the GPU 1 card from the rig and replaced it twice with 2 different GPUs (different risers, power cable...) on the PCI slot. They all got a lot of HW errors…
When they are recognized as GPU 2/3/4/5 they are working without issues.

I also leaved the PCI slot of GPU 1 empty and only pu the cards on the other PCI slots of the motherboard: GPU 0 / GPU 2 / GPU 3 / GPU 4 / GPU 5



But I suppose the computer does assign the PCI slot of GPU 2 (PCIEX1_2) as GPU slot 1. Indeed there are always HW errors indicated in GPU 1…while physically there is nothing on this slot.

I configured the BIOS as indicated in the manufacturer's instructions. (https://www.gigabyte.com/MicroSite/462/mining.html)

I can't figure out what is causing this issue.. ? A BIOS setup? A hardware problem with the motherboard? It's a GA-H110-D3A (Gigabyte). Bios is up to date.

I am using HiveOs and Teamredminer (But also tried under other miner and SimpleminingOs)

Thank you 🙂



Try only one thing, low the frecuency of the memory to 1900 mhz and see if the errors continue. If this fix the problem then is time to change the thermalpads, Thermalright ones are good and cheap.
member
Activity: 325
Merit: 42
Hello,

I have 5 RX580 8Gb cards under HiveOS.

GPU 1 is always getting HW errors... but the strange thing is that it's happening with any card listed on GPU 1 !



I first thought it came from the card or the riser but I figured out that this is not hardware related problem.

I thought it's the biosmod but I tried with the stock rom and it it does not change anything

I removed the GPU 1 card from the rig and replaced it twice with 2 different GPUs (different risers, power cable...) on the PCI slot. They all got a lot of HW errors…
When they are recognized as GPU 2/3/4/5 they are working without issues.

I also leaved the PCI slot of GPU 1 empty and only pu the cards on the other PCI slots of the motherboard: GPU 0 / GPU 2 / GPU 3 / GPU 4 / GPU 5



But I suppose the computer does assign the PCI slot of GPU 2 (PCIEX1_2) as GPU slot 1. Indeed there are always HW errors indicated in GPU 1…while physically there is nothing on this slot.

I configured the BIOS as indicated in the manufacturer's instructions. (https://www.gigabyte.com/MicroSite/462/mining.html)

I can't figure out what is causing this issue.. ? A BIOS setup? A hardware problem with the motherboard? It's a GA-H110-D3A (Gigabyte). Bios is up to date.

I am using HiveOs and Teamredminer (But also tried under other miner and SimpleminingOs)

Thank you 🙂


Looking at your picture of the mobo with risers it seems you have the first riser in the PCIe 16 slot and not in the PCIEX1_1 slot. The PCIe 16 slot should be empty.
In other words only use PCIEX1_1/2/3/4/5 slots.
legendary
Activity: 1834
Merit: 1136
I have such motherboards, but I had other problems. Your errors may be due to hardware problems on the motherboard. About the main problem is a lack of power or overclocking. But the lack of power or overclocking may be due to hardware problems on the motherboard.Check all power circuits, overclocking parameters.
newbie
Activity: 40
Merit: 0
Hello,

I have 5 RX580 8Gb cards under HiveOS.

GPU 1 is always getting HW errors... but the strange thing is that it's happening with any card listed on GPU 1 !

https://i.ibb.co/FxWwP8J/1.png

I first thought it came from the card or the riser but I figured out that this is not hardware related problem.

I thought it's the biosmod but I tried with the stock rom and it it does not change anything

I removed the GPU 1 card from the rig and replaced it twice with 2 different GPUs (different risers, power cable...) on the PCI slot. They all got a lot of HW errors…
When they are recognized as GPU 2/3/4/5 they are working without issues.

I also leaved the PCI slot of GPU 1 empty and only pu the cards on the other PCI slots of the motherboard: GPU 0 / GPU 2 / GPU 3 / GPU 4 / GPU 5

https://i.ibb.co/HKmjpVH/2.png

But I suppose the computer does assign the PCI slot of GPU 2 (PCIEX1_2) as GPU slot 1. Indeed there are always HW errors indicated in GPU 1…while physically there is nothing on this slot.

I configured the BIOS as indicated in the manufacturer's instructions. (https://www.gigabyte.com/MicroSite/462/mining.html)

I can't figure out what is causing this issue.. ? A BIOS setup? A hardware problem with the motherboard? It's a GA-H110-D3A (Gigabyte). Bios is up to date.

I am using HiveOs and Teamredminer (But also tried under other miner and SimpleminingOs)

Thank you 🙂


Have you tried to leave pcie 16x empty or just don use it? And use all pcie 1x wherever you wanna use and leave pcie 16x empty.
full member
Activity: 1424
Merit: 225
First what is GPU1? Who is calling it that? Device manager? miner?
Does it map directly to a physical slot or does it change depending on installed cards?
Is the problem related to one physical slot or does it change depending on installed cards?
Determine if the problem is a slot or a particular card.

My guess, and it's only a guess due to the confusing problem description, is you have a bad card.
member
Activity: 325
Merit: 42
Hi

If the GPU would have a failure, it would not be working fine without HW errors on another slot (as it do) I think Smiley


If you have an old videocard with 2/3 GB mem put that in the PCIe 16 slot and use that for just video output, you have to change
Initial Display Output to PCIe 1 Slot
Internal Graphics to disabled
and see if that works, don't forget to set the order of the mining cards in your mining software without the vga card.
newbie
Activity: 3
Merit: 0
Hi

If the GPU would have a failure, it would not be working fine without HW errors on another slot (as it do) I think Smiley
member
Activity: 1558
Merit: 69
Wow, you probably hit on of the eletronic / tech issues where there's no explanation  Cheesy

You did a good job already trying most common problems, bios update, change risers, GPUs etc

You can try to change some settings on bios for sure
I have another motherboard, but I had some issues with 3 cards, the motherboard only recognized 2, no matter which positions, so I change settings on Bios, specifically on PCI-e section, and after some try and errors I found the issue.

You can try to change 1 parameter at time and see how it goes
Another thing to try, put a monitor or a adaptor on this GPU1
Try to clean this slot with isopropyl alcohol or a clean contact product

After that, I don't have more ideas, let's see if some person here have some idea

Or, you slot could be broken, some contact or pci lane is damaged


Thank you for your reply Smiley

I tried mostly everything in the bios, on pcie and other parameters... nothing changed.
I cleaned the slot, but the problem is still there.

I don't think it's the slot itself. Like I showed on this picture :

https://ibb.co/7S7qNX1

I removed the riser from slot PCIEX1_1. I moved this card with the riser to slot PCIEX1_5.
On PCIEX1_5 there was no HW errors.

The card that was on PCIEX1_2 was running fine before I moved the PCIEX1_1 to PCIEX1_5.
After moving the PCIEX1_1 card, the PCIEX1_2 card was generating a lot of HW errors.

On the mining software, the PCIEX1_2 card was displayed as GPU1.

Very strange thing..

This GPU has a failure, there is nothing you can do. I have the same problem with 1x rx570. After 2 years it started with HW failures.
newbie
Activity: 3
Merit: 0
Wow, you probably hit on of the eletronic / tech issues where there's no explanation  Cheesy

You did a good job already trying most common problems, bios update, change risers, GPUs etc

You can try to change some settings on bios for sure
I have another motherboard, but I had some issues with 3 cards, the motherboard only recognized 2, no matter which positions, so I change settings on Bios, specifically on PCI-e section, and after some try and errors I found the issue.

You can try to change 1 parameter at time and see how it goes
Another thing to try, put a monitor or a adaptor on this GPU1
Try to clean this slot with isopropyl alcohol or a clean contact product

After that, I don't have more ideas, let's see if some person here have some idea

Or, you slot could be broken, some contact or pci lane is damaged


Thank you for your reply Smiley

I tried mostly everything in the bios, on pcie and other parameters... nothing changed.
I cleaned the slot, but the problem is still there.

I don't think it's the slot itself. Like I showed on this picture :

https://ibb.co/7S7qNX1

I removed the riser from slot PCIEX1_1. I moved this card with the riser to slot PCIEX1_5.
On PCIEX1_5 there was no HW errors.

The card that was on PCIEX1_2 was running fine before I moved the PCIEX1_1 to PCIEX1_5.
After moving the PCIEX1_1 card, the PCIEX1_2 card was generating a lot of HW errors.

On the mining software, the PCIEX1_2 card was displayed as GPU1.

Very strange thing..
legendary
Activity: 2450
Merit: 1472
Wow, you probably hit on of the eletronic / tech issues where there's no explanation  Cheesy

You did a good job already trying most common problems, bios update, change risers, GPUs etc

You can try to change some settings on bios for sure
I have another motherboard, but I had some issues with 3 cards, the motherboard only recognized 2, no matter which positions, so I change settings on Bios, specifically on PCI-e section, and after some try and errors I found the issue.

You can try to change 1 parameter at time and see how it goes
Another thing to try, put a monitor or a adaptor on this GPU1
Try to clean this slot with isopropyl alcohol or a clean contact product

After that, I don't have more ideas, let's see if some person here have some idea

Or, you slot could be broken, some contact or pci lane is damaged
newbie
Activity: 3
Merit: 0
Hello,

I have 5 RX580 8Gb cards under HiveOS.

GPU 1 is always getting HW errors... but the strange thing is that it's happening with any card listed on GPU 1 !

https://i.ibb.co/FxWwP8J/1.png

I first thought it came from the card or the riser but I figured out that this is not hardware related problem.

I thought it's the biosmod but I tried with the stock rom and it it does not change anything

I removed the GPU 1 card from the rig and replaced it twice with 2 different GPUs (different risers, power cable...) on the PCI slot. They all got a lot of HW errors…
When they are recognized as GPU 2/3/4/5 they are working without issues.

I also leaved the PCI slot of GPU 1 empty and only pu the cards on the other PCI slots of the motherboard: GPU 0 / GPU 2 / GPU 3 / GPU 4 / GPU 5

https://i.ibb.co/HKmjpVH/2.png

But I suppose the computer does assign the PCI slot of GPU 2 (PCIEX1_2) as GPU slot 1. Indeed there are always HW errors indicated in GPU 1…while physically there is nothing on this slot.

I configured the BIOS as indicated in the manufacturer's instructions. (https://www.gigabyte.com/MicroSite/462/mining.html)

I can't figure out what is causing this issue.. ? A BIOS setup? A hardware problem with the motherboard? It's a GA-H110-D3A (Gigabyte). Bios is up to date.

I am using HiveOs and Teamredminer (But also tried under other miner and SimpleminingOs)

Thank you 🙂
Jump to: