Support for using GPUs from different nodes on one job

Hi,
I know it's a lot to ask but it would be great to have GPUs from different clients working on one job.

I know that the skip and limit parameters can be used to split up cracking jobs but finding the right split points will be tough. Two ideas would be:

1. Bench the different systems and split the keyspace proportionately
2. Break down the keyspace into slices (for example 10 minute aequivalents) and distirbute them round robin.

Both solutions would only work if the key space is deterministic , with makes working on larger maskfiles or rulesets + wordlists very complicated.

Also the splitting up of chunks is highly dependent on the hashtype (10 min MD5 vs 10 min Keepass)

No files, tests, or entry points are named in the issue. Start by mapping how jobs, keyspace splitting, skip/limit parameters, and clients are represented; done would mean distributing one deterministic cracking job across GPUs on different clients while accounting for hashtype-dependent speeds.

Written by the indexing model from the issue text.