GPU/MIC Servers Our Products: Custom Rackmount Servers > GPU/MIC Servers

Servers featuring the Intel Xeon® Phi™ (MIC) Coprocessor The highly anticipated Intel Xeon Phi is now shipping and Aspen Systems offers a large selection of Phi-optimized, high performance systems specifically designed to utilize and expand all of its amazing capabilities and infinite number of new, complex project capabilities. Intel's Xeon Phi is an immensely powerful, parallel-processing, co-processor which opens up a world of new high-performance computing opportunities for users in the fields of research, discovery, mathematics, weather, science and more. When combined with Intel's E5-2600 series processors, the Xeon Phi puts Exascale-level computing within the budgets and range of previously ignored scientific users. The Intel Xeon Phi is based on 22-nanometer technology and utilizes 3D tri-gate transistors while using only 225 watts of power. With up to 8GB of GDDR5 memory, the Phi 5110P model delivers up to 1.01 Teraflops of double-precision performance. Prominent research and supercomputing labs have already incorporated Intel's Xeon Phi Coprocessors into their latest super computers. As the Xeon Phi becomes widely available, there is no reason to be left behind. Let Aspen Systems help you with your high-performance computing goals. NVIDIA® Tesla® Kepler K10/K20/K20X GPU Optimized Servers Aspen Systems configures builds and sells high-performance systems using the highly acclaimed Kepler K10, K20 and K20X GPU's from NVIDIA. The latest Kepler Core technology brings new possibilities to a wider range of highly technical applications that require lightning fast computational speed. Dynamic Parallelism makes it possible for GPU threads to dynamically spawn additional ones while allowing the GPU greater data adaptation. SMX Streaming multiprocessing, parallel programming and accelerated GPU performance is achieved through the most commonly used algorithms. In addition to their blistering fast speed, this exciting new Kepler Architecture is more energy-efficient than any GPU card ever made. Kepler GPUs are up to 3x more power efficient than previous GPU offerings from NVIDIA. By increasing the number of CUDA cores by 4-times, reduced clock speeds allow higher performance at lower energy costs. With tons of developers already jumping on the Kepler bandwagon, you can expect a broad range of new computing capabilities for building your next GPU server. Aspen Systems offers a wide-array of Kepler GPU-ready platforms including 1U, 2U, 3U and 4U models with 1 - 4 Kepler GPU's in a single box.

Filter Products:

Form Factor 1U 2U 2U Twin 4U 5U Processor 1x Intel Xeon E5-2600 2x AMD Opteron 6200 2x Intel Xeon 5600 2x Intel Xeon E5-2600 8x Intel Xeon 7500 Max Memory 192GB 256GB 512GB 1024GB # Hard Drives 3 4 6 8 16 Features GPU/MIC Ready Infiniband QDR

Displaying 8 of 8 Systems Form Factor Processor Max. Memory Hard Drives Key Features SuperServer SYS-1017GR-TF Configured at: $1,729.50 1U Single Socket Intel® Xeon® E5-2600 Series Up to 256GB DDR3 ECC Reg. (6) 2.5" SATA/SAS Hot-swap Up to 4 x NVIDA Tesla GPU and Intel Xeon Phi Cards SuperServer SYS-1027GR-TRF Configured at: $2,680.00 1U Dual Socket Intel® Xeon® E5-2600 Series Up to 256GB DDR3 ECC Reg. (4) 2.5" SATA/SAS Hot-swap Up to 3 x Tesla GPU and Intel Xeon Phi Redundant Power Supplies SuperServer SYS-2027GR-TRF Configured at: $2,855.25 2U Dual Socket Intel® Xeon® E5-2600 Series Up to 256GB DDR3 ECC Reg. (10) 2.5" SATA/SAS Hot-swap Up to 4 x NVIDIA Tesla GPU and Intel Xeon Phi Redundant Power Supplies SuperServer SYS-7047GR-TPRF Configured at: $2,680.00 4U Dual Socket Intel® Xeon® E5-2600 Series Up to 512GB DDR3 ECC Reg. (8) 3.5" SATA/SAS Hot-swap Up to 4 x Tesla GPU and Intel Xeon Phi Redundant Power Supplies Alpine G122I Configured at: $5,910.00 1U Dual Socket Intel® Xeon® 5600 Series Up to 192GB DDR3 ECC Reg. (3) 3.5" SATA/SAS Hot-swap (2) NVIDIA Tesla GPUs in a 1U chassis Alpine G122A Configured at: $6,045.00 1U Dual Socket AMD Opteron™ 6100/6200 Series Up to 256GB DDR3 ECC Reg. (3) 3.5" SATA/SAS Hot-swap (2) NVIDIA Tesla GPUs in a 1U chassis Mesa G221I Configured at: $3,028.00 2U Twin Dual Socket Intel® Xeon® 5600 Series (per node) Up to 192GB DDR3 ECC Reg. (per node) (6) 3.5" SATA/SAS Hot-swap (per node) (1) NVIDIA Tesla GPU (per node) 2 Nodes in a 2U chassis Alpine G424I Configured at: $2,305.00 4U Dual Socket Intel® Xeon® 5600 Series Up to 192GB DDR3 ECC Reg. (8) 3.5" SATA/SAS Hot-swap (4) NVIDA Tesla GPU Cards in a 4U chassis No systems found that match your filter selections Looking for a system that isn't available online? Our experienced sales engineers are standing by and can design a high performance computing solution that fits your exact needs. Call us at (888) 227-1286 or email us at sales@aspensystemsdirect.com for free design assistance.

GPUs are revolutionizing the way scientists, mathematicians and engineers are able to sort, process and analyze large, complex sets of data. Since the name was coined by NVIDIA just a few years ago, Graphics Processing computing has taken off and quickly becoming accepted worldwide by millions of high-end workstation users around the globe. Even the most advanced and powerful CPU's of today are limited in making user applications run faster. GPUs give unprecedented power to high-end applications by taking the computational-intensive duties away from the CPU. The mutual benefit is that traditional computing resources are free to continue normal processing tasks while the GPU handles the more complex, floating point calculations of user application code. While today's advanced CPU's are running on multiple cores, GPUs can include up to thousands of their own, smaller, more efficient cores. The CPU runs the serial portion of the application code while the GPU handles the parallel. It's the perfect tandem.

GPUs – Programmable Development Platforms

While a GPU can be as much as 100-time faster than a CPU for complex application computations, the key to unlocking that performance lies in the graphics programming capabilities. Today's GPU architecture supports the latest CUDA parallel programming model, allowing users to maximize their application performance by tapping into the resources of the larger, scientific community. Early on, NVIDIA understood the importance of making GPUs fully programmable for the scientific and research fields. The other benefit is how NVIDIA has been able to continue bringing to market faster, more advanced GPUs with literally no down-time from users associated with their application development. GPU programming is a natural transition for developers of common C, C++ and Fortran languages. Thousands upon thousands of universities and research labs are advancing the latest scientific breakthroughs through parallel, GPU-computing and its vast community of developers.

GPUs – a Powerful High Performance Server Combination

As a proud partner of NVIDIA, Aspen Systems is able to offer the latest GPU / GPU Server technology based on the Tesla platforms. Together with NVIDIA, Aspen System's GPU Server solutions are the culmination of three decade's worth of expertise developing and designing high performance computing solutions in the HPC industry. GPUs represents a natural fit for our expertise in building high-performance, enterprise servers. A GPU Server is a high performance computing server able to perform high performance computations. By utilizing the power of today's multi-core, enterprise server with the latest, high-performance GPUs, there are literally thousands of cores working simultaneously to meet the challenges of the most demanding computing requirements. But the benefits of the GPU Server doesn\u2019t end there: A GPU Server is a green solution. Offloading the tasks of a CPU-only solution will reduce heat, lower energy costs and drastically improve power consumption in server rooms and data centers. The GPU server is a versatile and green computing solution.

GPUs – High Performance Computing Solutions by Aspen Systems

As a leader in the high-performance, HPC computing industry, Aspen Systems can provide supercomputing GPUs or a GPU Server integrated with NVIDIA\u2019s latest TESLA architecture and bundle them for accomplishing your specific tasks. Whether your needs are scientific modeling, advanced mathematics applications, computational finance, weather modeling or data analysis, Aspen Systems has the engineering, design and production experience to produce state-of-the-art, GPUs to get the job done.