Closed UserDC-LeGrand closed 1 year ago
Is there a use case for this? Do we need arguments for each device type to have it's own set of algorithms?
I would say it’s a must when mining CPU and GPU on same rigs. Case in point: I wanted to point my GPU’s at etchash and CPU at ghostrider, so I used the -Algorithm ethash,gr in my flightsheet. What SWARM then pointed both my GPU’s and CPU to ghostrider, instead of Ethhas for GPUS, and gr for the CPU. Perhaps new SWARM arguments could be implemtend such as Algorithm_GPU and Algorithm_CPU to support Algo allocation to proper devices.
just thought of something. I could try running two instances of SWARM. My flightsheet could have a custom miner for the CPU, and one for the GPUs only,. Worst case, the Cpu miner I select, would not profit switch...
I don't know if that will work. I have never tried it. I will look into changing -Algorithm to -GPUAlgorithm and -CPUAlgorithm respectively if it is needed.
I made a fork from this issue.
On that fork I:
1.) Removed -Algorithm argument. 2.) Added -GPUAlgorithm1 argument. 3.) Added -GPUAlgorithm2 argument. 4.) Added -GPUAlgorithm3 argument. 5.) Added -CPUAlgorithm argument.
Obviously the -GPUAlgorithm argument corresponds to device group (NVIDIA1,NVIDIA2,NVIDIA3)
If you want to stress test it and report, please do.
If no issues found, I will release as is...But I will still need to add all arguments and options to get help
command
I want to be able to send specific Algo to GPU, and to CPU, with -Algorithm argument. Both GPU and CPU default to same algo ...