Added 2 RX580 GPUS to a 6 RX 580 RIG, now Windows menus open slow if at all.

[miner_in_bkk]

Hi Guys,

Have a frustrating problem, which I think I understand the cause of, but cant figure out how to fix it.

Biostar 250 BTC Pro (12 GPU slots)
6 MSI RX 580 8gb
4gb Ram 2133
Celeron 3900
Claymore 10.5
Virtual Memory setting 16384
Windows 10 x64 Pro
AMD Blockchain Driver

I have just received another 6 Gigabyte RX 580 8gb Gaming GPU’s and tried to add 2 GPU’s to the 6 GPU’s and see how problematic this would be before trying to add the full amount of the 6 GPU’s to the one rig.

Order of events.

1. Device Manager saw the 8 – Great.
2. Windows installed automatically the Block Chain Driver. – PATCHED - REBOOT
3. Claymore saw the 8 – Great.
4. Claymore got random GPU “dag creation” errors. - FIXED the error, as I saw a post stating to increase Virtual Memory to 64000. – REBOOT
5. NOTICED THE PROBLEM. Windows takes 15 seconds to open any folder / AMD Radeaon App / Windows Setting / Start Menu wont open. - REBOOT
6. Problem still there. Open Claymore and get 20-22 MHS on the 2 Cards (30.8 on the other 6). Remembered to modd the GPU’s with at least the timing straps @ 1500– did it. But cant flash it – Cant open the CMD with administration rights due to not having access via start menu.
7. Stopped my other rig, and removed a 4gb stick of 2133 ram to add into this rig.
8. Display Menu is now working fine. Flash the Bios. Patcher says nothing to patch. – Reboot.
9. Claymore now getting 30.5-30.8 Mhs with all 8. – YAHOO.
10. Leave it for 30 mins to watch the Miner, now the PROBLEM has come back. All menus / setting menu’s take FOREVER to open up. Could be 2 minutes with several attempts.
11. Tried changing the Virtual Memory but no setting with reboot solves this problem.
12. Removed the two GPU’s, took out the 2nd 4gb RAM and took the Virtual Memory back to 16384 - Rebooted and the problem has now gone.

Allot of miners have success with 10-12 GPU rigs with 4gb. And the problem exists if I have 8gb in the rig, then maybe its not the physical RAM, even though the problem momentarily went away when i added the additional 4gb.

Any ideas ?

[philipma1957]

do you use msi afterburner with this setup?

it polls all the gpus and the cpu every 1000 microseconds  same as 1 time each second

6 cards and a cpu  is a lot of data   60 times per minute


set the poll speed to 11000  it will help.

go to settings
click monitoring

set from  1000 to 11000

[mini miner]

is windows 10 is new version? manytimes blockchain driver is not working good with new win10 os. try new driver 18.2.1. it will help u.i have same problem when i installed new driver it solved.

[miner_in_bkk]

do you use msi afterburner with this setup?

it polls all the gpus and the cpu every 1000 microseconds  same as 1 time each second

6 cards and a cpu  is a lot of data   60 times per minute


set the poll speed to 11000  it will help.

go to settings
click monitoring

set from  1000 to 11000

Any undervaluing / overclocking is done via Claymore. I dont use MSI AB.

[miner_in_bkk]

is windows 10 is new version? manytimes blockchain driver is not working good with new win10 os. try new driver 18.2.1. it will help u.i have same problem when i installed new driver it solved.

I am using Windows 10 Version 1709.

I downloaded the Adrenaline drivers, but i haven't used them. Just the Blockchain version. I will do a DDU then try with adrenaline.

Many thanks

[tipac]

you want to install 8 gpu with only 4 gb of ram ? wtf
you didn't mention if you plug hdmi in motherboard or in gpu ?
first thing to do: check if you have any problem in safe mode. if no problem in safe mode then it means its driver problem, in this case use ddu then install fresh copy of driver and you shouldn't have any problem

[miner_in_bkk]

you want to install 8 gpu with only 4 gb of ram ? wtf
you didn't mention if you plug hdmi in motherboard or in gpu ?
first thing to do: check if you have any problem in safe mode. if no problem in safe mode then it means its driver problem, in this case use ddu then install fresh copy of driver and you shouldn't have any problem

I was wondering the same with the 4gb of RAM. I have read quite a few posts, where 4gb does 12 GPU's, so this was my first stop. But i knew it couldn't be the RAM as when i added another 4gb the problem still existed.

Anyway, i managed to overcome it. I added one by one, increased the Virtual Memory to 24000. And so far it has run for 12 hours no problem at all with 4gb.

I did find a slightly loose 8 PIN connection on one of the GPU's. So maybe that was the reason ?

Still using the block chain driver too.

Will report back once i add the other 4.

[Random Seller]

I don’t think ram will ever be an issue with mining.

Can you open task manager and check if disk is at 100% all the time?

And are you mining on the cpu as well?

[bigjee]

Any undervaluing / overclocking is done via Claymore. I dont use MSI AB.

Did you verify overclock to make sure there are no errors.
My system becomes laggy when i push overclocks on a card too far.

[miner_in_bkk]

I don’t think ram will ever be an issue with mining.

Can you open task manager and check if disk is at 100% all the time?

And are you mining on the cpu as well?

Just opened task manager, and the disk is 0-1%, CPU is 55%, and RAM was 2.2 out of 4.

The problem has gone for now, but i haven't added the other 4 GPU's yet.

I just checked for errors, and 1 card had 20,000 errors since 10 hours. Power went off 10 hours ago. The other 7 had 0,0,0,3,2,516,640.

I do plan to remove claymore and the command lines mclock, cclock, cvddc, mvddc, etc, and use OverdriveNtool with Phoenix miner.

Although, when the power went off earlier, only the first 6 cards were being used by Claymore and not the latest 2. Used remote restart via TeamViewer and still the same result. It wasn't until i got home and did a full shut down and turn on, that it picked up the 8 again. Any ideas ??

BTW, Just changed my second rig from claymore to OverdriveNtool with phoenix miner today and so far no errors, slightly better Mh's, slightly lower Watt at the wall meter. So happy about that.

[lncm]

Try newest adrenalin drivers.

When set to "compute" the performance is the same or better than august blockchain drivers, and much more stable...

[solosequenosenada]

Try newest adrenalin drivers.

When set to "compute" the performance is the same or better than august blockchain drivers, and much more stable...

I have never had problems with the blockchain drivers of August, months running the RIG without reboots or stability problems. I have tested the new adrenaline drivers and there is no performance increase at least with my RX470 RIG

[crocozino]

Try newest adrenalin drivers.

When set to "compute" the performance is the same or better than august blockchain drivers, and much more stable...

I have never had problems with the blockchain drivers of August, months running the RIG without reboots or stability problems. I have tested the new adrenaline drivers and there is no performance increase at least with my RX470 RIG

same here, I've been using blockchain drivers for mixed rig, which is consist of one RX VEGA, 3x RX480 and 2xRX570
and the only problem I have - the voltage not applying for all cards, but if change some clocks settings in afterburner - then the correct voltage settings are been applied same moment. veird