|

GPU accelerators and coprocessors for PowerEdge servers

PowerEdge GPU Coprocessor
Accelerators are available for the PowerEdge R720T620 and C8220x servers and the C410x PCIe expansion chassis.

Increase the performance of your PowerEdge data center.

Extract some of the highest levels of performance from your Dell PowerEdge servers through a general-purpose computation on graphics processing units (GPUs) architecture. When you add GPU processing power to the CPU capabilities already available in your PowerEdge servers, you open the door to outstanding performance across hundreds of processing cores.

  • GPUs are high-performance, many-core processors that can be used to accelerate a wide range of applications.
  • Advanced GPU programming methods and toolkits enable easy integration into your data center.
  • GPU processors can be internally installed in standard PCIe slots or connected externally via host interface cards (HICs) through C410x PCIe expansion chassis.
  • To take full advantage of GPU technology, applications need to be programmed and compiled for GPUs using NVIDIA® CUDA® or AMD with OpenCL™ toolkits.

 

Use GPU computing to accelerate applications and enable large-data processing in many types of applications, including:
• Analytics
• Computer-aided design (CAD)
• Data-intensive imaging
• Engineering
• Forecasting
• Life sciences
• Media, entertainment and animation
• Multiapplication environments
• Research

                       Poweredge GPU 
What is GPU computing?

GPU computing is the use of a GPU as a coprocessor in tandem with a CPU to accelerate CPU intensive applications typically used in scientific and engineering domains. GPUs are also being used in virtual desktop environments.

The GPU accelerates applications running on the CPU by off-loading some of the compute-intensive and time-consuming portions of the application code. The rest of the application still runs on the CPU. From a user’s perspective, the application may run faster because it’s using the massively parallel processing power of the GPU to boost performance. This is known as “heterogeneous” or “hybrid” computing.

In the 1999–2000 time frame, computer scientists, along with researchers in fields such as medical imaging and electromagnetics, started using GPUs to accelerate a range of scientific applications. This was the advent of the movement called GPGPU or general purpose GPU computing.

A CPU consists of four to sixteen CPU cores, while a GPU consists of hundreds of smaller cores. Together, they operate to crunch through the data in the application. This massively parallel architecture is what gives the GPU its high compute performance. There are a number of GPU-accelerated applications that can provide easy access high-performance computing (HPC).

Card Option

AMD

AMD FirePro™ S9000/S7000/W7000
• Focused on virtual device interface (VDI) and
  visualization environments
• Double-wide form factor (S9000)
• Single-wide form factor (S7000/W7000)
• Up to 225W and over 1200 cores
• PCIe Gen 3
AMD FirePro S9000/S7000/W7000
AMD FirePro V7800P/V7800
• Focused on VDI and visualization environments
• Double-wide form factor
• Up to 138W and over 1400 cores
• PCIe Gen 2
AMD Firepro V7800P/V7800
Dell does not support local DVI/Display Port connections on Graphics cards. These cards are designed to be used as VDI devices and support virtualized graphics.

 GPU for Virtualization and Graphics — AMD
BrandAMD Firepro™AMD FireproAMD Firepro AMD Firepro AMD Firepro
ModelV7800PV7800S9000S7000W7000
Dell ServersPE R720PE T620PE R720PE R720PE T620
Cores/Processors14401440179212801280
Memory2GB GDDR52GB GDDR56GB GDDR54GB GDDR54GB GDDR5
Max Power Consumption138W138W225W150W150W
Single Precision2.0 TFLOPS2.0 TFLOPS3.23 TFLOPS2.4 TFOPS2.4 TFOPS
Dual Precision400 GFLOPS400 GFLOPS806 GFLOPS152 GFLOPS152 GFLOPS
CoolingPassiveActivePassivePassiveActive
WidthSingle WideSingle WideDouble WideSingle WideSingle Wide
LengthFull LengthFull LengthFull LengthFull LengthFull Length
HeightFull HeightFull HeightFull HeightFull HeightFull Height
*above GPU options may also be utilized in compute environments


 Coprocessors for Compute — Intel®
BrandIntel® Xeon Phi™Intel® Xeon Phi™Intel® Xeon Phi™
Model7120P5110P3120A
Dell ServersPE R720
PE C8220X
PE R720
PE C410X
C8220X
PE T620
Cores/Processors616057
Memory*16GB
GDDR5
8GB
GDDR5
6GB
GDDR5
Max Power Consumption300W225W300W
Single Precision2.44
TFLOPS
2.022
TFLOPS
2.006
TFLOPS
Dual Precision1.220
TFLOPS
1.053
TFLOPS
1.003
TFLOPS
CoolingPassivePassiveActive
WidthDouble WideDouble WideDouble Wide
LengthFull LengthFull LengthFull Length
HeightFull HeightFull HeightFull Height


 GPU for Compute — NVIDIA 
BrandNVIDIA TeslaNVIDIA TeslaNVIDIA TeslaNVIDIA TeslaNVIDIA TeslaNVIDIA TeslaNVIDIA Tesla
ModelM2075C2075M2090K10K20MK20CK20X
Dell ServersPE R720
PE C410X
PE T620PE R720
PE C8220X
PE C410X
PE R720PE R720
PE C8220X
PE T620PE R720
Cores/Processors4484485123072249624962688
Memory6GB GDDR56GB GDDR56GB GDDR58GB GDDR55GB GDDR55GB GDDR56GB GDDR5
Max Power Consumption200W215W225W225W225W225W235W
Single Precision1.03 TFLOPS1.03 TFLOPS1.33 TFLOPS4.58 TFLOPS3.52 TFLOPS3.52 TFLOPS3.95 TFLOPS
Dual Precision515 GFLOPS515 GFLOPS665 GFLOPS190 GFLOPS1.17 TFLOPS1.17 TFLOPS1.31
TFLOPS
CoolingPassiveActivePassivePassivePassiveActivePassive
WidthDouble WideDouble WideDouble WideDouble WideDouble WideDouble WideDouble Wide
LengthFull LengthFull LengthFull LengthFull LengthFull LengthFull LengthFull Length
HeightFull HeightFull HeightFull HeightFull HeightFull HeightFull HeightFull Height




  GPU for Virtualization and Graphics — NVIDIA
BrandNVIDIA TeslaNVIDIA Quadro®NVIDIA Quadro®NVIDIA Quadro®NVIDIA GRIDNVIDIA GRID
ModelM2070QQ6000K2000K4000K1K2
Dell ServersPE R720
PE C410X
PE T620

PE R720
PE T620

PE R720
PE T620

PE R720PE R720
Cores/Processors4484483847687683072
Memory6GB GDDR56GB GDDR52GB GDDR53GB GDDR516GB GDDR38GB GDDR5
Max Power Consumption225W204W51W80W130W225W
Single Precision1.03 TFLOPS1.03 TFLOPS733 GFLOPS1.246 TFLOPSNANA
Dual Precision515 GFLOPS515 GFLOPS NANA
CoolingPassiveActiveActiveActivePassivePassive
WidthDouble WideDouble WideSingle WideSingle WideDouble WideDouble Wide
LengthFull LengthFull LengthFull LengthFull LengthFull LengthFull Length
HeightFull HeightFull HeightFull HeightFull HeightFull HeightFull Height


Dell HPC Solutions

At the core of Dell's high-performance computing (HPC) solution is a modular HPC infrastructure stack built from industry-standard hardware components and best-of-breed partner products. These solutions offer the performance and scalability required by the most demanding applications at a fraction of the cost of proprietary systems.

With partners in each of the key sections of an HPC solution, we can design your solution through a building-block strategy that’s tested and validated for performance and reliability. Below are the key building blocks:

• Compute Platform — Dell rack-mounted servers, blades and shared-infrastructure systems
• Processors and accelerators
• Operating system
• Provisioning and system management software
• Middleware and development tools — Dell offers flexibility with the choice of standards-based message passing libraries, including MPICH, MPICH-GM and MVAPICH, supporting available interconnect options. Dell and its partners also offer a broad selection of comprehensive development toolkits, including compilers, math libraries and system and development tools.
• Interconnect technologies
• Storage

Learn more at Dell HPC Solutions.
Desktop Virtualization Solutions (DVS) Enterprise

Accelerate remote access to 3D virtualized graphics applications with Desktop Virtualization Solutions (DVS) Enterprise. Designers, engineers and business professionals can be more productive and highly collaborative across the enterprise through accelerated access to multimedia and graphics-intensive applications from virtually anywhere, on virtually any device. DVS Enterprise connects and accelerates business by allowing your team to view, display and interact with high-end graphics and multimedia content reliably and securely.