The Colfax CXT6000 series clusters democratizes HPC by providing scientists, engineers and technical professionals with an affordable supercomputing resource, that is much faster, uses fewer systems and is more energy-efficient than a CPU-only driven cluster. Powered by the NVIDIA® Tesla™ S1070 Computing System, a four-teraflop 1U system, coupled with NVIDIA CUDA™ architecture, Colfax CXT6000 series clusters provide a quantum leap in performance and continue to accelerate the pace of engineering and scientific work to solve the most computationally-
“The Tesla GPU-based CXT6000 series clusters by Colfax are enabling researchers with supercomputing performance from smaller clusters at a significantly lower cost, greatly improving the scope and size of the research problems that they can solve, or the number of simulations they can complete per day,” said Andy Keane, general manager of the Tesla business at NVIDIA. “More importantly, they can do this while consuming less electricity, reducing the carbon footprint required to power these systems and cool the data centers that house them.”
The new Colfax CXT6000 series includes:
1. Each S1070 1U GPU Computing System has four Tesla T10 GPUs, delivering almost four teraflops of performance.
2. 4GB of memory per GPU, support for IEEE 754 single & double floating point precision, and a fast 102 GB/sec GDDR3 interface to the memory, the S1070 1U speeds the transition to energy-efficient parallel computing and scales to solve the world’s most important computing challenges more quickly and accurately.
• CXT6000N, CXT6000H, CXT6000HS, CXT6000E Computing Nodes.
1. CXT6000N includes twin 1U compute nodes with latest Intel® Xeon® 5500 series "Nehalem" processors and up to 96GB RAM per node.
2. CXT6000HS includes twin 1U compute nodes with Intel® Xeon® 5400 series processors and up to 128GB RAM per node.
3. CXT6000H includes 2U compute nodes with Intel® Xeon® 5400 series processors and up to 128GB RAM per node.
4. CXT6000E includes twin 1U compute nodes with latest AMD Opteron™ "Shanghai" processors and up to 128GB RAM per node.
• High-speed 40Gb/s InfiniBand connectivity from Mellanox Technologies to deliver low-latency and high-bandwidth for performance-
Colfax’s pre-configured and delivered ready-to-go Tesla GPU-based, InfiniBand accelerated clusters enable customers to quickly and easily leverage the power of clusters to accelerate business and research outcomes. The available solutions come pre-configured in 4, 8, 16-node options that deliver 16, 32, 64 teraflops of performance respectively, or may be customized to suit individual customer requirements.
“As datasets grow increasingly complex, the path to great discovery gets more and more computationally challenging,”
“The combination of GPUs and high-speed 40Gb/s InfiniBand connectivity provides new level of performance and compute speeds that are essential in order to solve the most complex engineering and scientific problems,” said Gilad Shainer, director of technical marketing at Mellanox Technologies. “Colfax’s custom architecture utilizes the highest productivity compute and interconnect components to deliver efficient, tailored solutions for our mutual customers.”
The Colfax CXT6000 cluster is immediately available. All the models can be fully customizable to suit individual customer requirements. For more information please visit: http://www.colfax-
# # #
About Colfax International
Colfax International is a global provider of customized workstations, servers, clusters and storage solutions. Founded in 1987, Colfax International is based in Sunnyvale, California and is privately held. For more information, please visit www.colfax-intl.com.