PDC

PDC

Below follows a short description of the PDC systems. For more information regarding the systems and available software on the systems, please see the PDC web pages http://www.pdc.kth.se/

Lindgren, Cray XES, Capability system for massively parallel workloads

The Cray XE6 is a capability system consisting of 1516 dual 12-core notes connected by a 3D-torus Gemini network.  The CPUs are AMD Opteron 6100 series 2.1 GHz.  The theoretical peak performance (TPP) is 305.6 TFlop/s and the total amount of memory is 47.3 TByte.  This machine is dedicated to large scale, wide jobs (i.e. massively parallel jobs); typical job size should be in the thousands of cores and preferably growing larger over subsequent allocation rounds. Large allocations with small jobs will waste the significant investments made to allow efficient execution of wider jobs. Large allocations with jobs smaller than a few hundred cores will not be able to run efficiently.

 

Mass storage
A number of storage options including HSM are available at PDC. Each user has a private storage area in the HSM system, hsm.pdc.kth.se