Desktop RAM limits

You are allocated 8 GB of RAM per user/RE session. Therefore we recommend that you only use the local environment for "lightweight" research or exploratory work, and the HPC for "heavyweight" batch research.

How will this impact my session?

Capping at 8 GB of memory should not impact most work. However, R and Rstudio place a substantial drain on the memory usage. You may find such queries run slowly, or may crash. We recommend running such queries on the HPC.

Will this also affect my HPC usage?

No, not all. We recommend working on the HPC wherever possible.

Can I monitor the memory usage myself?

You can see the top five applications that are using the most memory with the following command:

ps -o pid,user,%``<span class="c-mrkdwn__highlight">mem``,%cpu,command ax | sort -b -k3 -r | head -n 6

The output provides a snapshot of its current memory usage in the form of a percentage of 12GB.

Managing memory usage in R

When using R you can manage your memory usage by:

1. Remove any unused large objects with rm(<object_name>)
2. After object removal, run gc()

Removing an object in R will not actually free the memory that it used. Therefore running the gc() command in R will (largely) return the memory to the system.