How about within the miner program running on machine with multiple mining devices (ASIC, GPU, etc), does the work gets divided down? I would suppose so right?
Yes. Most pools use an extraNonce2 size of 4 bytes which gives each miner 2^32 different block headers that can be created from a single workload. The mining program can than split all this work up to all the different devices connected to it.
Thanks GoldenWings91!
I was posting this related scenario on other forum:
1 remote wallet connected by 1 machine with 2 GPUs doing solo mining
vs
1 remote wallet connected by 2 machines with 1 GPU each doing solo ming.
So the first case I would have higher chance of finding block compared to the 2nd case since wallet does not function like a pool where work is dished out ... based on your reply, looks like I am right?