TACC provides the computational user community access to terascale computing systems, remote and collaborative visualization resources, high-performance, local disk file systems, and a petascale data archive.
The Sun Constellation cluster, Ranger, enables users to scale applications to a petascale level. Batch queues allow jobs in the tens of thousands of cores that have access to terabytes of distributed memory and parallel file systems that can accommodate files in the terabyte range. Programming models supported include MPI, OpenMP, and the combination of both (hybrid paradigm). Configuration details and information on how to efficiently use the system can be found in the Ranger User Guide.
The Dell Linux Xeon cluster, Lonestar, should be the computational resource of choice for users with parallel, MPI codes that take advantage of a large, distributed-memory architecture. Development and serial queues enable users to port and develop models that can scale from tens, to hundreds, or even thousands of processors. See the Lonestar User Guide for system configuration and instructions on how to efficiently use the system.
The IBM AIX Power5 system, Champion, provides users access to a large, shared memory (SMP) system. Codes that require large memory or employ OpenMP to accomplish parallelism on large memory nodes will benefit from the use of this system. MPI codes can also be executed to leverage the power of multiple of the Power5 nodes in a small-scale, parallel mode. A parallel GPFS file system provides a solution to users with a requirement for high-speed IO. See the Champion User Guide for system configuration and instructions on how to efficiently use the system.
The Dell Serial cluster, Stampede, accommodates users with a requirement to solve serial and small-scale parallel problems. Users have access to a large, parallel work file system for temporary storage and the TACC petascale storage facility for storing files on a permanent basis. This system also provides the community with a development platform to begin scaling algorithms to eventually run on the larger TACC compute clusters. See the Stampede User Guide for system configuration and information on how to use the system.
The TACC Visualization Laboratory contains several hardware and software resources dedicated to providing interactive, immersive, and collaborative visualization environments. In addition Spur is available for the remote visualization of large data sets.