There are five main queues/partitions on SCIAMA:

• sciama2.q : This queue comprises of 95 nodes each with 16 cores and 64 Gbytes of shared memory. These nodes are named node100 to node194 .
  
• sciama3.q : This queue comprises of 48 nodes each with 20 cores and 64 Gbytes of shared memory. These nodes are named node200 to node247 .
• sciama4.q : This queue comprises of 12 nodes each with 32 cores and 192 Gbytes of shared memory. These nodes are named node300 to node311.
• sciama4-12.q : This queue comprises of 16 nodes each with 32 cores and 384 Gbytes of shared memory. These nodes are named node312 to node327.
• himem.q : There is only one node in this queue called vhmem01 . This node is a high memory node with 16 cores and 512 Gbytes of shared memory. Only use this queue if your job requires over 384 Gbytes of memory.
• hicpu.q : There is only one node in this queue, node350 which has 128 cores and 125Gbybtes of memory.  Ideal for multiple short jobs that doesn’t require much memory.

Additionally, there are a few more specialist queues:

• jupyter.q : This queue comprises of a subset of sciama3.q nodes and is used by default to run jupyter notebook servers on SCIAMA.
  
• training.q : This queue comprises of a small subset of nodes and should be used for any of the training exercises (to limit interference with the rest of SCIAMA).
•  rsm1.q: This queue comprises of a small subset of the sciama2.q and is used by RemoteSolverManager software by a dedicated user.