Disadvantages if starting miner for each GPU separate?

[mensa84]

Hello,

do you know if I will have disadvantages if I start the miner program (in my case ethminer) for each GPU separate?
Like starting on one rig that commands:

Code:

ethminer.exe --cuda-devices 0 ...
ethminer.exe --cuda-devices 1 ...
ethminer.exe --cuda-devices 2 ...
ethminer.exe --cuda-devices 3 ...

and so on ...

The advantage of that solution will be, that if one GPU fails (for example if overclocked too much) only that single GPU will not mine until problem is fixed. If I start all GPUs in one process and any of the GPUs fail, the mining will stop for all GPUs until problem is fixed.

But will that have any negative (or even positive) impact on mining performance or anything else?
Thanks for your help.

[car1999]

interesting idea, did you try it?

I think multiple instances requires more CPU and memory, if you have I7 CPU and 16G memory, that's fine.

[mensa84]

Yes, I am trying it since some hours to find out max. OC settings for each GPU to run stable (cause I have cards from different models/manufacturers on that rig, but is also interesting for same model/manufacturer as I found out now).

I have only a Celeron G3900 2.80 GHz and 4 GB RAM on that rig and it seems to be good.
I am not sure if it does really need more CPU/RAM. Yes there are more processes in that case, but maybe smaller processes instead of one big process.

An interesting thing would also be to name each card as an own worker. For example Rig1.GPU1, Rig1.GPU2, Rig1.GPU3, etc.

But I am still unsure, if that does have any other impacts.

[cashen]

I really like the idea. One other downside would be remote monitoring using Awesome Miner. Would need to bind a separate port for each API.

How is it working? DO you notice a overall increase or decrease?

[mensa84]

The effective hashrate at ethermine stays the same. But I don't have the technical background, if that would bring up any else negative things.

I think I will let it run like that until every GPU ran stable for about 1 week nearly at max. possible OC. Then I will switch back to only one mining process per rig. The target should be stability and then I won't need that anymore. I hope.

[cashen]

The effective hashrate at ethermine stays the same. But I don't have the technical background, if that would bring up any else negative things.

I think I will let it run like that until every GPU ran stable for about 1 week nearly at max. possible OC. Then I will switch back to only one mining process per rig. The target should be stability and then I won't need that anymore. I hope.

Might want to try one of these to help manage the windows

http://windowsreport.com/tabbed-command-line-windows-10/

[hyet24]

this is very useful when i first started.  i would have save tons of time finding the bad gpu for the first week mining.

[mensa84]

Thx, but I think Windows has this built in since version 7 or so?
Right click on task bar ->