White Papers and other technical publications written by staff in the Research Computing group.
|Kernel Control in a High Performance Compute Cluster|
We wanted a plug and play mechanism for installing and maintaining the software for accelerator devices, such as NVidia Tesla cards, into our HPC cluster nodes. At first glance this sounds like a simple task, but the frequent system updates applied by our configuration management system can lie dormant on a node until such time as it has been rebooted, and thus play havoc with system version detection in automated installers.
This white paper by Adam Carrgilson (CiS) describes a method for maintaining the device's kernel modules (LKM) in step with the current installed kernel and kernel headers, by triggering a node reboot and an automatic compilation of the drivers.
|An Introduction to Data Centres at NBI|
Supercomputers, server ‘farms’ and large-scale data stores are an essential infrastructure underpinning modern biological science. For reasons of cost, complexity and practicality it is not feasible to host these systems in a scientific laboratory.
This white paper by Paul Fretter (CiS) outlines some of the basic design features of a data centre and the reasoning behind it.