hello, i'm wondering how to run cgminer for many workers. i want to run it for 6 gpu's (r9 290). how many workers do i need to run it smooth?
let's try it with this example config:
cgminer --scrypt -o stratum+tcp://pool:port -u user.worker -p workerpassword -I 13 --auto-fan --thread-concurrency 28456 --gpu-engine 950 --gpu-memclock 1250
now:
1) is it possible to run multiple workers in one cgminer config?
2) if not, do i have to run 6 times cgminer with different setup?
in this case looks like each cgminer will be controling each gpu and i don't like this idea too much, so:
2a) how to set up fans and other parameters like tc, gpu-engine - do i have to set them for each cgminer/gpu manually or they can stay like this?
2b) how should it look like if i had for example 4x r9 290 + 2x r9 290x
please help me answer this questions, i bet it can help lots of people
You have to know the order that cgminer will identify your gpu's. Lets assume they showed up like this:
GPU0: 290X
GPU1: 290
GPU2: 290X
GPU3: 290X
GPU4: 290X
GPU5: 290
If, for instance, you wanted intensity = 13 for your 290x's and 15 for your 290's (just for this example) you would use something like this
cgminer --scrypt -o stratum+tcp://pool:port -u user.worker -p workerpassword -I 13,15,13,13,13,15 --auto-fan --thread-concurrency 28456 --gpu-engine 950 --gpu-memclock 1250
This holds true for other parameters. Simply separate them by commas in the order that the GPU's are identified. Any parameters that contain only 1 statement - like "--auto-fan" - will be applied to all cards.
As for your question about using a worker for each GPU, I'm not sure as I've never tried this. I use a unique worker for each rig so I can tell if a machine has gone down or not. I don't see the benefit of using a worker for each GPU. Sounds like a lot of extra overhead.
Hope this helps.