The Alabama Supercomputer Authority houses the SGI UV and DMC supercomputers.

SGI UV (Ultraviolet)

Image of SGI UV 2000 Supercomputer

The SGI UV at the Alabama Supercomputer Center has 268 CPU cores, 4160 GB of shared memory, and 182 terabytes in the GPFS storage cluster.

The compute system is configured with 256 Xeon E5-4640 CPU cores operating at 2.4 GHz and 4 TB of shared memory. The front end node is configured with 12 Xeon E5-2667 CPU cores operating at 2.9 GHz and 64 GB of memory. This gives the entire system a floating point performance of 5194 GigaFLOPS.

 

 

 

UV 2000 Datasheet  |  Silicon Graphics, Inc.

 

Dense Memory Cluster (DMC)

The DMC at the Alabama Supercomputer Center has 2216 CPU cores and 16 terabytes of distributed memory. Each compute node has a local disk (up to 1.9 terabytes of which are accessible as /tmp). Also attached to the DMC is a high performance GPFS storage cluster, which has 45 terabytes of high performance storage accessible as /scratch from each node. Home directories as well as third party applications use a separate GPFS volume and share 137 terabytes of storage.

The machine is physically configured as a cluster of 8, 16, 20, or 36 CPU core SMP boards. Four nodes have 2.26 GHz Intel quad-core Nehalem processors and 24 gigabytes of memory. Four nodes have 2.3 GHz AMD 8-core Opteron Magny-Cours processors and 32 gigabytes of memory. Forty nodes have 2.5 GHz Intel 10-core Xeon Ivy Bridge processors and 128 gigabytes of memory. One node has a 2.3 GHz Intel 12-core Haswell processors and 12 gigabytes of memory.  One node has a 1.3 GHz Intel 64-core Knights Landing preproduction processor and 94 gigabytes of memory. Twelve nodes have 2.1 GHz Intel 18-core Broadwell processors and 128 gigabytes of memory.

The DMC has 32 NVIDIA GPU (Graphic Processing Unit) chips. These are a combination of: two Tesla S1070 units (external GPUs connected in pairs to four DMC

nodes), four DMC nodes configured with two Tesla M2070 cards each, and four DMC nodes configured with four Tesla K20m cards each. These multicore GPU chips are similar to those in video cards, but are installed as math coprocessors. This can give significant performance advantages for software that has been adapted to use these processors. Thus the processing capacity of the DMC cluster is: 

Conventional processing capacity - 31.24 TFLOPs 
Single precision GPU capacity - 74.92 TFLOPs 
Double precision GPU capacity - 23.53 TFLOPs