I know the question is old, but I was asked today if I knew the answer of this unanswered question.
You could use the -p
along with the --machinefile
options:
Julia can be started in parallel mode with either the -p
or the --machine-file
options. -p
n will launch an additional n worker processes, while --machine-file
file will launch a worker for each line in file file. The machines defined in file must be accessible via a password-less ssh login, with Julia installed at the same location as the current host. Each machine definition takes the form [count*][user@]host[:port] [bind_addr[:port]]
. user defaults to current user, port to the standard ssh port. count
is the number of workers to spawn on the node, and defaults to 1. The optional bind-to bind_addr[:port]
specifies the IP address and port that other workers should use to connect to this worker.
It has been a long time since I used the --machinefile
option, in my case the n
option didn't work and I don't know if it has been fixed, but you could add one line for each worker process you want instead, for example, if this doesn't work for you:
# machinefile.txt
23 user@host
You could try this:
# machinfile.txt
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
user@host
And then invoke julia like:
$ julia -p 2 --machinefile machinefile.txt
For a total of 25 processes (2 local and 23 remote).
But the n
option should work if it is documented, else please check if there is a bug and if not, open a new one.