By default a R.E.D. Server Development job is executed on a node with two quad-cores Intel® Xeon® E5430 processor (i. e. 8 cores) and 64 GB RAM. Such computer harware is generally sufficient for quantum mechanics (QM) computation involving small molecules. However for a big molecular system and/or when using a large QM basis set the computational time dramatically increases. Thus to limit this time problem a user can request to work on a more powerful node, i. e. to use up to two ten-cores IntelĀ® Xeon E5-2690 v2 processor (i. e. 20 cores) and up to 256 GB RAM.

'node0' to 'node13' contains 64 GB RAM, while 'node14' to 'node19' incorporates only 16 GB. The user has now the possibility to force the call of a node with 64 GB RAM. To choose a node open the qstat interface to look at which node is free or running, and provide the "KeyWordNode$n" keyword (where '$n' is a node number between 0 and 19). By default the 'node0' → 'node13' are first called if free. Thus, forcing the call of a node with 64 GB RAM is only useful if the 'node0' - 'node13' are all busy.

To be able to execute her/his job(s) on a powerful node a user has to contact by email the q4md-forcefield tools team (at contact_at_q4md-forcefieldtools.org; _at_ = @) to describe her/his project and computational needs. If the request is found justified, the user will received by email a 'temporary-job related keyword' to activate the use of such a 20 cores node. The user provides this keyword at the "SUBMISSION OF DATA 1/5" step during the job submission procedure (read also FAQ 17).

Simply leave the provided "DEFAULT" keyword to use the default node type in your R.E.D. Server Development job.