Calc, our standalone Linux computation server, was released to the campus almost 5 years ago to the day (August 16, 2013). It has served Carnegie relatively well over those 5 years for small scale computation with relatively high memory requirements. However, it's starting to show its age. The time has come for Calc to retire. Once Calc goes offline for the SRCF shutdown, it will not return.
And in it's place shall rise a new Calc, the Calc vGPU Cluster!
The new Calc will be just like the old Calc, where code can be run within certain limits (8GB memory per user max) and with long running codes auto-reniced, but with a GPU added for graphical apps and GPGPU codes. In addition, 4 more nodes with the same power as Calc will be available for jobs that need additional, dedicated CPU, Memory, and if necessary GPU resources.
This new Calc cluster is also intended to provide a resource for DPB and DGE that can act as a training ground for users that are new to large scale computation. As users computational needs progress, transitioning from on-demand code execution to resource requisition tends to be a challenge. Now that the new Calc provides both traditional code execution capabilities as well as resource requisition capabilities, it is the ideal resource where learn the ins and outs of job submission and resource requisition before utilizing larger computing resources like Memex.