The amount of memory available on the interactive nodes is described on the Interactive (login) Nodes page. Since the interactive nodes are shared among all users care should be taken not to be disruptive. One way to do this is to try to use too much memory and if you think your work might be disruptive you should either run it in batch or run it interactively on a batch node as described here.
The memory on the batch nodes is a consumable resource in SGE, which means that SGE knows how much memory each node has and how much memory each job requires. SGE assumes by default that your job requires 1.1GB but you can specify that your job requires more or less memory than this according to the instructions in Submitting Jobs. However, it should be noted that the more memory you require the harder it will be for SGE to schedule your job, especially when the cluster is filled to capacity. Therefore it is in your best interest to not require memory more memory than you need. Also, note that different batch nodes had different amounts of memory per job slot as described at Compute Nodes.