Try our new documentation site (beta).
Client for a compute server
You have two options for indicating that a Gurobi program will act as a client of a compute server. If you are writing a program that calls the Gurobi C, C++, Java, .NET, or Python API's, these API's provide routines that allow you to specify the names of the compute servers. If you use these routines, then Gurobi licenses aren't required on the client.
Alternatively, you can set up a gurobi.lic
file that points to
the compute server. This option allows you to use a compute server
with nearly any program that calls Gurobi, without the need to modify
the calling program. The client gurobi.lic
file should
contain a line that looks like this:
COMPUTESERVER=machine1.mydomain.com,machine2.mydomain.com,machine3.mydomain.comor like this:
COMPUTESERVER=192.168.1.100,192.168.1.101,192,168.1.102This line provides a comma-separated list of Gurobi compute servers. If your compute servers use a password, you should also include a line that gives the password:
PASSWORD=cspwd
If you'd like to specify a job priority, you can add a line that gives an integer priority value:
PRIORITY=7Higher priority jobs run before lower priority jobs. Please consult the Gurobi Compute Server section of the Gurobi Reference Manual for more information.
Note that if your client and server are both running on the same
machine, you can add a COMPUTESERVER=localhost
line to your
gurobi.lic
file.
The compute server will ignore this line, and
the client will ignore everything but this line. Another option in
this situation is to create a separate gurobi.lic
file for the
client, and to set the GRB_LICENSE_FILE
environment variable
to point to this file.