We've unified Citrix solutions and our portfolio. Learn more .


Citrix Ready Product Image / Logo

The new generation of GPU computing server PR2764GW is a newly introduced 2U dual-rack GPU accelerated computing server.

Request Product Details

Compatible with

  • XenServer 6.5


  • Citrix Hypervisor (XenServer)

Product Details

The new generation of GPU computing server PR2764GW is a newly introduced 2U dual-rack GPU accelerated computing server. Using Intel C612 high-performance chipset, support Intel E5-2600V3 series processors, the model has 16 DDR4 DIMM slots up to support memory capacity 1TB, support 10 hot-swappable 2.5-inch hard drive, support for four full-length full-length double Wide GPU card slot, the other provides a PCIE 3.0 x8 slot, supports SATA / SAS, onboard integrated 2 1Gb network interface, supports a variety of different network options: 10Gb Ethernet, 40Gb, 56Gb InfiniBand, can be enhanced High-speed performance and I/O flexibility to meet the interconnection needs of different applications; onboard integrated BMC, support IPMI 2.0 remote management, 2000W platinum redundant power; new generation PR2764GW high-density GPU accelerated computing server main applications and high-performance computing (HPC), the GPU takes on part of the heavy budget and time-consuming code that accelerates applications running on the CPU, helping you do more computing tasks, work with larger datasets, and reduce application up-time.

Application Environment

High performance computing (HPC), data mining, and large data analysis

PR2764W is suitable for the mainstream 32-bit and 64-bit HPC applications, mainly for computational physics, computational materials, computational chemistry, life sciences, gene, protein structure research, pharmaceutical engineering, petroleum exploration, satellite signal processing, CAD/CFD structural mechanical fluids Mechanical simulation, weather forecast and other traditional areas of HPC, in the Internet, finance, data mining, machine learning, large data analysis and other emerging areas of high-performance computing has gradually been widely used.

As a new application field of high performance computing, Deep Learning is a hotspot of machine learning in recent years. It has made a breakthrough in image recognition and speech recognition. Its application pattern is a combination of large data and deep neural network model. Parallelization of data or deep network models with GPU clustering speeds up program execution. Using GPU to accelerate the depth of learning, training depth learning network, you can give full play to thousands of GPU computing core of the efficient parallel computing capacity in the use of massive data training data scenarios, the time-consuming significantly reduced, less occupied servers. The use of GPU clusters as the infrastructure to build depth learning / machine learning platform, has become the preferred solution in the field, and widely used in the Internet industry.

GPU in the field of high-performance computing, making the CPU no longer the only choice for computing chips. Compared with the CPU, GPU has more powerful computing power (current Nvidia K80 with 4992 core, double-precision floating-point computing power up to 2.9TFlops, internal storage bandwidth 480GB/s), the task processing mode is more simple, gradually applied High-performance computing in all areas, help the rapid development of the industry. GPU with its powerful computing power, to attract users to use GPU to accelerate the efficiency of application execution, while GPU also has the advantages of low cost, high performance, low power consumption, reduce the user's total cost of ownership.