Supercomputing Laboratory - Home

Skip to Main Content

Colorado School of Mines

CSM Supercomputing Laboratory


» Supercomputing Laboratory » Resources


Resources


Thanks to the generous support of our sponsors and partners, members of the Supercomputing Laboratory have access to the following high performance computing resources:

Jaguar at Oak Ridge National Laboratory (Jaguar has been ranked No. 1 in the 34th TOP500 Supercomputer List)

  • 224,256 cores in 18,688 nodes
    • Each node has two Opteron 2435 "Istanbul" processors linked with dual HyperTransport connections
    • Each processor has six cores with a clock rate of 2600 MHz supporting 4 floating-point operations per clock period per core
  • 299 TB of memory
    • Each node is a dual-socket, twelve-core node with 16 gigabytes of shared memory
    • Each processor has directly attached 8 gigabytes of DDR2-800 memory
  • 2331 Teraflops (1759 Teraflops on Linkpack)
    • Each node has a peak processing performance of 124.8 gigaflops
    • Each core has a peak processing performance of 10.4 gigaflops
  • 3D torus interconnection network

Kraken at the University of Tennessee (Kraken has been ranked No. 3 in the 34th TOP500 Supercomputer List)

  • 99,072 cores in 8,256 nodes
    • Each node has two Opteron 2435 "Istanbul" processors linked with dual HyperTransport connections
    • Each processor has six cores with a clock rate of 2600 MHz supporting 4 floating-point operations per clock period per core
  • 129 TB of memory
    • Each node is a dual-socket, twelve-core node with 16 gigabytes of shared memory
    • Each processor has directly attached 8 gigabytes of DDR2-800 memory
  • 1028.85 Teraflops (831.7 Teraflops on Linkpack)
    • Each node has a peak processing performance of 124.8 gigaflops
    • Each core has a peak processing performance of 10.4 gigaflops
  • 3D torus interconnection network

Ranger at the University of Texas (Ranger has been ranked No. 9 in the 34th TOP500 Supercomputer List)

  • 62,976 cores in 3,936 nodes
    • Each node has four AMD Opteron Quad-Core 64-bit processors (16 cores in all) on a single board, as an SMP unit
    • Each processor has four cores with a clock rate of 2300 MHz supporting 4 floating-point operations per clock period per core
  • 123 TB of memory
    • Each node is a 16-core SMP uint with 32 GB of memory
  • 579 Teraflops (433.2 Teraflops on Linkpack)
    • Each node has a peak processing performance of 147.2 gigaflops
    • Each core has a peak processing performance of 9.2 gigaflops

  • The interconnect topology is a 7-stage, full-CLOS fat tree with two large Sun InfiniBand Datacenter switches at the core of the fabric (each switch can support up to a maximum of 3,456 SDR InfiniBand ports)

Ra at Colorado School of Mines (Ra has been ranked No. 134 in the 31st TOP500 Supercomputer List)

  • 2,144 cores in 268 nodes
    • 256 nodes with two 512 Clovertown E5355 quad-core processor at a clock rate of 2670 MHz supporting 4 floating-point operations per clock period per core
    • 12 nodes with four 48 Xeon 7140M dual-core processor at a clock rate of 3400 MHz supporting 4 floating-point operations per clock period per core
  • 5.6 TB of memory
    • 16 GB memory per node for 184 nodes
    • 32 GB memory per node for 72 nodes
  • 23.2 Teraflops (17 Teraflops on Linkpack)
    • Clovertown node has a peak processing performance of 85.44 gigaflops
    • Xeon node has a peak processing performance of 108.8 gigaflops
  • Cisco SFS 7024 IB Server Switch

Alamode Lab at Colorado School of Mines

  • 94 cores in 24 nodes
    • Each node has two dual-core AMD Opteron 2218 processors at a clock rate of 2600 MHz supporting 2 floating-point operations per clock period per core
  • 192 GB of memory
    • Each node has 8 gigabytes of memory
  • 488.8 Gigaflops peak performance
    • Each core has a peak processing performance of 5.2 gigaflops
  • Gigabit Ethernet Switch

Microsoft Donated Cluster (MSC) at Colorado School of Mines

  • 32 cores in 16 nodes
    • Each node has a dual-core AMD Opteron 2212 processor at a clock rate of 2000 MHz supporting 2 floating-point operations per clock period per core
  • 128 GB of memory
    • Each node has 8 gigabytes of memory
  • 128 Gigaflops peak performance
    • Each core has a peak processing performance of 4 gigaflops
  • Gigabit Ethernet Switch

Nvidia Tesla S1070 system cuda1.mines.edu at Colorado School of Mines

  • This 1U rack mountable system contains 4 of the Nvidia Quadro FX 5600 GPU cards and has a peak computing performance of 4 trillion floating point operations per second, or 4 Teraflops.
  • Each of the 4 graphics processing units (GPU) on the Tesla has 240 processing cores and 4 Gbytes of memory for a total of 960 cores and 16 Gbytes.
  • The individual GPUs are connected to the front-end node via a PCI connector.