The insatiable demand for processing power is often at odds with challenges, such as limited data center space and power. General purpose computing on graphics processing units (GPGPU) can help you break down those barriers.
GPU solutions pack more computing power into the same space — and use less energy per calculation — by adding the number-crunching horsepower of graphics processing chips to the equation. Partner with Dell to bring a new standard of performance to your HPC cluster:
- Quicker time to results — GPUs repurposed for HPC applications offer massive increases in performance, decreasing time to results.
- More efficient space and energy use — GPU solutions use less space and energy than traditional HPC installations. GPU solutions either install into standard PCIe slots inside systems or connect externally as a dedicated server or PCIe chassis using a host interface card.
- Better reliability and scalability — HPC software development aimed at GPUs (often using the NVIDIA® CUDA® or AMD OpenCL™ toolkit) has been heavily tested and validated by the open-source community, ensuring scalability and reliability.
- Greater system agility — The flexible configuration of GPU solutions (installed in standard PCIe slots or externally as a dedicated server or PCIe chassis) gives you the ability to build new HPC solutions or add compute power to existing HPC infrastructures, balancing the advantages of CPU and GPU computing.
- A wider range of options — At the core of our GPU-accelerated solutions is a stack of world-class software, hardware and services that have been tested, validated and tuned for performance, so you can take full advantage of your new compute power faster.
With the addition of accelerated technology servers to the Dell HPC portfolio, you have access to broader adoption models to expand performance for an existing infrastructure or implement new systems optimized for maximum density, memory and throughput.
Featured Case Studies

National Center for Supercomputing Applications (NCSA)
Learn how Dell empowers research at the National Center for Supercomputing Applications.
Recommended Software
Learn about the range of development options available for general-purpose computing on graphics processing units.

gpgpu.org — General-Purpose Computation on Graphics Hardware
Get the latest news and information about general-purpose computing on graphical processing units (GPGPU) at this central resource for the GPGPU community.

Khronos Group OpenCL
Access a range of resources for OpenCL (Open Computing Language) programming, the first open, royalty-free standard for cross-platform, parallel programming of modern processors from Khronos Group, the creators of OpenCL.

AMD OpenCL
Get resources from AMD for OpenCL (Open Computing Language) programming. Preserve your expensive source code investment and easily target both multicore CPUs and the latest graphics processing units (GPUs), such as those from AMD.

AMD Developer Forum: GPU Developer Tools
Visit an insightful forum where graphics processing unit (GPU) developers can ask questions, share challenges and offer the latest news and solutions in a friendly, thought-provoking atmosphere.

NVIDIA CUDA Zone
Dig into a wealth of information about CUDA, NVIDIA's parallel-computing architecture that enables dramatic increases in computing performance by harnessing the power of graphics processing units (GPUs).

Fermi: NVIDIA’s Next-Generation CUDA Compute Architecture
Get the facts on Fermi, NVIDIA's next-generation CUDA computing and graphics architecture.

Video: Scott Morton of Hess Talks About CUDA
Scott Morton talks about Hess Corporation's experience with CUDA and the advantages of accelerating seismic processing with graphics processing unit (GPU) computing.
Blade Server Systems

PowerEdge M610x Blade Server
With PCI Express (PCIe) expansion options and a feature-rich Chassis Management Controller, the Dell PowerEdge M610x blade server enables you to efficiently run applications, consolidate your data center and simplify data management.

PowerEdge M1000e Modular Blade Enclosure
Built to combat data center sprawl and IT complexity, the Dell PowerEdge M1000e delivers one of the most energy-efficient, flexible and manageable blade server product on the market.
Rack Systems


PowerEdge C6145 Rack Server
Get fast answers to the most demanding high-performance computing (HPC) problems with the PowerEdge C6145, the only server in its class with two 4-socket AMD Opteron 6100 processor-based servers in 2U.

PowerEdge C6100 Rack Server
In a shared infrastructure with four two-socket server nodes in a 2U rack chassis, the PowerEdge C6100 server is designed to minimize environmental impact and provide hyperscale capabilities in HPCC, virtual and cloud environments.


PowerEdge C410x: Paving the Way for Powerful Thinkers
Take a closer look at the Dell PowerEdge C410x, an external PCI Express (PCIe) expansion chassis supporting server connections to up to 16 graphics processing unit (GPU) cards, enabling massive parallel calculations separate from the server.

Expanding the boundaries of GPU computing
Supporting up to 16 PCI Express devices in a flexible, highly efficient design, the Dell PowerEdge C410x expansion chassis helps organizations take advantage of the next step in high-performance computing architectures — GPU computing.
Technical Resources
Get the most out of your investment using our tactical guidance and best practices guides.

NCSA Answers Questions About GPU Computing
Get the basic facts about using graphics processing units (GPUs) for high-performance parallel computation. These high-performance 'many-core' processors can accelerate a wide range of science and engineering applications.

GPU Acceleration of Equations Assembly in Finite Elements Method
Explore a method for mapping the equations assembly problem for St. Venant-Kirchhoff material to a GPU computation model that can be considered as a general technique to manage complex composition with the different requirements of GPU resources.

GPU Acceleration of Molecular Modeling Applications
Learn how graphics processing units (GPUs) are being applied to molecular modeling computation throughout the scientific world.




