PRIMERGY CX270 S1 is a dual socket server node for ambitious High Performance Computing, Analytics and Visualization solutions.

Request Information

Compatibility

  • down-arrow Citrix DaaS
    Citrix DaaS

Product Details

The PRIMERGY CX270 S1 condensed half-wide form factor optimizes large data center scale-out deployment, resulting in less rack space used and lower power envelopes as compared to standard rack servers.

Each half-wide, 2U high server node with Intel® Xeon® processor E5-2600 family CPUs, 16 DIMM slots, 3 PCIe Gen3 I/O slots features an optional GPGPU accelerator card, chosen from parallel processing NVIDIA® Tesla™ 20 series, enabling a ten-fold performance boost in HPC applications. Moreover, using shared power supplies and cooling fans, 

Two server nodes can be embedded in the PRIMERGY CX400 enclosure, only using 2U space in conventional standard datacenter racks.

Features

  • Two CX270 S1 server nodes, each with dual processors, up to 512 GB memory optional high-end NVIDIA GPGPU local storage capacity of up to 6 drives can be smartly packaged into a condensed 2U rack enclosure, using shared cooling and power.
  • 50% less rack space used as compared to equivalently configured standard rack servers.
  • Double the server density results in more performance achieved per rack unit. With GPGPU support, up to 10 times faster execution per server node in HPC computing is achieved.
  • CX270 S1 server nodes come w/o local fans or power supplies. Instead they share central cooling fans and hot plug power supplies per 2 U chassis.
  • Decreased energy consumption, lower investment, yet still redundant operation. Lower energy budgets for a comparable performance as with standard rack servers.
  • 4 to 8 core processors and Turbo Boost 2.0 PCIe Gen 3 support 16 DIMMs per server node with up to 512 GB DDR3 memory
  • Up to 70% performance improvement over Xeon 56xx series, more depending on applications.
  • 60 % improvement in I/O bandwidth for heavy load I/O communication. Its flexible memory dimension fits to any application requirement.
  • 1 optional GPGPU compute module per CX270 server node, selected from the market NVIDIA® Tesla™ 20 series, the parallel computing processors for high performance computing (HPC). Tesla M2075 or Tesla M2090 High-end GPGPU with 515 / 665 Gigaflops peak floating point performance double precision.
  • Tesla GPU’s high performance makes them ideal for seismic processing, biochemistry simulations, weather and climate modeling, signal processing, computational finance, CAE, CFD, data analytics and other high performance computing. Combined with new density optimized CX270 server nodes, it optimizes large data center scale-out deployment for less rack space used and lower power envelopes.

Screenshots

screenshot