GPU Accelerators for Dell EMC – Amulet Hotkey

Amulet Hotkey CoreModule
Powerful GPU Acceleration for
Dell EMC PowerEdge MX

GPUs for Dell EMC PowerEdge MX

Unique Amulet Hotkey CoreModule enables powerful GPU acceleration for MX7000.
Ideal for: graphics accelerated virtual desktops, virtual workstations, machine and deep learning, modelling, simulation, accelerated databases analytics, visualization and more.

Powerful Performance

Enhance user experience, run visualizations faster, accelerate demanding and evolving workloads

Uncompromised Efficiency

GPUs fit within the MX7000 chassis for superior reliability, performance and to maintain modular density

Flexible Scalability

Power virtual workloads and enhance software defined data center capabilities

Explore Amulet Hotkey CoreModule Solution

Solution Overview

Dell EMC PowerEdge MX7000 Modular Chassis

A 7U chassis with eight front bays to support a variety of compute and storage sled combinations

Amulet Hotkey CoreModule

A core expansion module that enables graphics and compute acceleration within the MX7000 chassis

Dell EMC PowerEdge MX740c Compute Sled

2-socket, single-width sled with flexible and dense compute resources for virtual workloads

Amulet Hotkey CoreCartridge

A compact cartridge with 2x NVIDIA Tesla T4 GPUs for software-defined GPU acceleration

Dell EMC PowerEdge MX840c Compute Sled

4-socket, double-width sled with elastic compute resources for mission-critical applications and performance workloads

Amulet Hotkey CoreModule Mezzanine Card

Dedicated PCIe 3.0 for reliable, fast and low-latency connection between the compute sled and CoreModule

Compare Solution Components

Product Description Benefits Notes
CoreModule

Download Datasheet
  • A core expansion module that enables GPU accelerators in Dell EMC MX7000
  • Fit up to 2x CoreModules in MX7000 Fabric B slots for up to 16x GPUs per chassis
  • Supports up to 4x CoreCartridges
  • Maintains modular architecture and density benefits
  • Fits within MX7000 chassis in fabric B slots for superior reliability and performance without sacrificing front bays
  • Advanced engineering and thermal management to support powerful GPU acceleration
  • Fully integrated into MX7000 chassis thermal management
  • Each CoreModule provides connection for one GPU to each of the MX7000 front eight bays
  • Use 2x CoreModules for up to two GPUs per MX740c
  • Use 2x CoreModules for up to 4x GPUs per MX840c
CoreCartridge
  • Compact GPU cartridge with 2x NVIDIA Tesla T4 GPUs
  • Fits in Amulet Hotkey CoreModule
  • Compact design that combines 2x NVIDIA Tesla T4 GPUs
  • Advanced thermal management
  • Pay-as-you grow by adding CoreCartridges to scale as required

Supports standard capabilities of the Tesla T4 including:

  • Virtualization support including VMware and Citrix
  • NVIDIA GRID Software
  • NVIDIA Quadro vDWS Software
  • NVIDIA Tesla software for HPC applications
CoreModule Mezzanine Card
  • Amulet Hotkey Mezzanine Card provides a dedicated, low-latency PCIe 3.0 connection between the compute sled and CoreModule
  • Designed to fit PowerEdge MX740c and MX840c compute sled Fabric B slot
  • Orthogonal connector for reliable, low-latency and high-speed PCIe 3.0 interconnect
  • Upgraded orthogonal connector with increased pin density to support greater power delivery for GPUs
  • For use with Amulet Hotkey CoreModules only
  • MX740c Fabric B requires dual CPU configuration
  • MX840c Fabric B requires four CPU configuration

Key GPU Accelerated Workloads

Use GPU acceleration to support new and challenging workloads across the organization. Deploy in private, public or hybrid cloud environments that require performance, agility and flexibility. MX7000 servers with CoreModules combine Intel Xeon Scalable processor and NVIDIA Tesla GPU accelerator performance with the flexibility, scalability and efficiency of the MX7000 platform for:

Enterprises are moving their users to modern, graphics-intensive Microsoft Windows 10 environments. Organizations that are using virtual desktop infrastructure (VDI) to support this migration need increasingly powerful server resources to support their VDI users

The Dell EMC PowerEdge MX7000 Modular Chassis, equipped with up to sixteen NVIDIA Tesla T4 GPUs contained in Amulet Hotkey CoreModules brings high-density graphics acceleration to the virtual enterprise environment.

Key factors:

  • Enhance user experience for users with modern productivity applications, multiple monitors and Windows 10
  • Offload server CPUs using virtual GPU with hardware graphics and display rendering
  • VMware and Citrix virtualization
  • NVIDIA GRID vPC and vAPP software
  • Reduce power consumption, comparing two Tesla T4 GPUs and one Tesla M10 that each provide the same user density and 32GB of video memory
  • Enhanced remote display protocol and video processing with H.264 and H.265 encode/decode acceleration

Enterprises are looking to achieve the same efficiency and operational benefits of VDI for their 3D, graphics-intensive and high-performance workstation applications. Organizations that are using virtual workstations to support this transition need increasingly powerful server resources to support their professional design and creative users.

The Dell EMC PowerEdge MX7000 Modular Chassis, equipped with up to sixteen NVIDIA Tesla T4 GPUs contained in Amulet Hotkey CoreModules brings a balance of high-speed processors and high-performance graphics and compute acceleration for virtual workstations.

Key factors:

  • NVIDIA Turing architecture fuses real-time ray tracing, AI, simulation and rasterization
  • NVIDIA RT Cores to accelerate real-time ray-tracing and to speed rendering workloads
  • Professional visualization and photorealism for workflows with large datasets and complex models, including applications infused with AI and machine learning technologies
  • VMware or Citrix virtualization
  • NVIDIA Quadro vDWS Software for certified workstation applications

Enterprises are deploying artificial intelligence with underlying technologies such as machine learning and deep learning by either active development of learning models or by simply using applications that incorporate machine learning technologies. Machine and Deep Learning workloads require high performance and considerable capacity to rapidly manage and analyse structured and unstructured data, and then deliver consistent and fast result that enable valuable insights for productivity and competitive advantage.

The Dell EMC PowerEdge MX7000 Modular Chassis, equipped with up to sixteen NVIDIA Tesla T4 GPUs contained in Amulet Hotkey CoreModules brings high-density virtual GPU acceleration for Machine and Deep Learning inferencing and training workloads.

Key factors:

  • NVIDIA Turing architecture with CUDA and Turing Tensor Cores
  • Breakthrough performance for FP32, FP16 to INT8 and INT4 precision computing
  • Support for AI and DL learning frameworks
  • NVIDIA Quadro vDWS software enables virtualized learning including VMware and Citrix environments
  • Bare-metal deployment using NVIDIA Tesla drivers and software

Enterprise success is increasingly tied to the ability to extract value from data using machine learning and GPU-accelerated databases to gain critical insights and competitive advantage. Advanced analytic workloads require high performance and considerable capacity to rapidly ingest, analyse and visualize structured and unstructured data, and then deliver fast and meaningful results.

The Dell EMC PowerEdge MX7000 Modular Chassis, equipped with up to sixteen NVIDIA Tesla T4 GPUs contained in Amulet Hotkey CoreModules brings Tensor Core flexibility for data science and data analytic workloads.

Key factors:

  • NVIDIA Turing architecture with Tensor Cores to accelerate end-to-end data science workflows
  • Tesla T4 universal AI accelerator for data-driven visualization and machine learning
  • GPU-database solutions to leverage the massive parallel processing capability of GPUs including real-time processing to tap into the enormous value of insights from live data streams
  • Advanced SDKs and libraries including CUDA-X AI with RAPIDS as well as broad industry frameworks
  • NVIDIA Quadro vDWS software to virtualize workloads in VMware and Citrix environments
  • Bare-metal deployment using NVIDIA Tesla drivers and software

Modern HPC data centers are key to solving some of the most important scientific and engineering challenges. HPC workloads require considerable compute and storage capacity to rapidly vast amounts of data to innovate and reduce the time to make new discoveries.

The Dell EMC PowerEdge MX7000 Modular Chassis, equipped with up to sixteen NVIDIA Tesla T4 GPUs contained in Amulet Hotkey CoreModules brings high-efficiency GPU acceleration for distributed computing environments running advanced research, simulation and modelling workloads.

Key factors:

  • High-efficiency scale-out performance that is ideal for distributed computing environments running single and mixed precision workloads
  • Delivers the latest NVIDIA Turing architecture and features including new Tensor Cores
  • Modern cloud paradigm using software-defined infrastructure flexibility to cover bare-metal, virtual and hybrid cloud workflows
  • NVIDIA Quadro vDWS software to virtualize applications in VMware and Citrix environments
  • Advanced SDK and libraries with broad industry framework and application support
  • GPU acceleration for over 600 HPC applications including ones infused with machine and deep learning technologies

Virtual Desktop and App Delivery

Enterprise are moving their users to modern, graphics-intensive Microsoft Windows 10 environments and organization supporting that migration with virtual desktop infrastructure (VDI) need increasingly powerful server resources to support their VDI users.

The Dell EMC PowerEdge MX7000 Modular Chassis, equipped with up to sixteen NVIDIA Tesla T4 GPUs contained in Amulet Hotkey CoreModules brings high-density graphics acceleration to the virtual enterprise environment.

Key factors:

  • Enhance user experience for users with modern productivity applications, multiple monitors and Windows 10
  • Offload server CPUs using virtual GPU with hardware graphics and display rendering
  • VMware and Citrix virtualization
  • NVIDIA GRID vPC and vAPP software
  • Reduce power consumption, comparing two Tesla T4 GPUs and one Tesla M10 that each provide the same user density and 32GB of video memory
  • Enhanced remote display protocol and video processing with H.264 and H.265 encode/decode acceleration

Virtual Workstation

Enterprise are looking to achieve the same efficiency and operational benefits of VDI for the 3D, graphics-intensive and high-performance workstation applications, and organization supporting transitioning to virtual workstations need increasingly powerful server resources to support their professional design and creative users.

The Dell EMC PowerEdge MX7000 Modular Chassis, equipped with up to sixteen NVIDIA Tesla T4 GPUs contained in Amulet Hotkey CoreModules brings a balance of high-speed processors and high-performance graphics and compute acceleration for virtual workstations.

Key factors:

  • NVIDIA Turing architecture fuses real-time ray tracing, AI, simulation and rasterization
  • NVIDIA RT Cores to accelerate real-time ray-tracing and to speed rendering workloads
  • Professional visualization and photorealism for workflows with large datasets and complex models, including applications infused with AI and machine learning technologies
  • VMware or Citrix virtualization
  • NVIDIA Quadro vDWS Software for certified workstation applications

AI, Machine and Deep Learning

Enterprise are deploying artificial intelligence with underlying technologies such as machine learning and deep learning by either active development of learning models or by simply using applications that incorporate machine learning technologies. Machine and Deep Learning workloads require high performance and considerable capacity to rapidly manage and analyse structured and unstructured data, and then deliver consistent and fast result that enable valuable insights for productivity and competitive advantage.

The Dell EMC PowerEdge MX7000 Modular Chassis, equipped with up to sixteen NVIDIA Tesla T4 GPUs contained in Amulet Hotkey CoreModules brings high-density virtual GPU acceleration for Machine and Deep Learning inferencing and training workloads.

Key factors:

  • NVIDIA Turing architecture with CUDA and Turing Tensor Cores
  • Breakthrough performance for FP32, FP16 to INT8 and INT4 precision computing
  • Support for AI and DL learning frameworks
  • NVIDIA Quadro vDWS software enables virtualized learning including VMware and Citrix environments
  • Bare-metal deployment using NVIDIA Tesla drivers and software

Data Analytics

Enterprise success is increasingly tied to the ability to extract value from data using machine learning and GPU-accelerated databases to gain critical insights and competitive advantage. Advanced analytic workloads require high performance and considerable capacity to rapidly ingest, analyse and visualize structured and unstructured data, and then deliver fast and meaningful results.

The Dell EMC PowerEdge MX7000 Modular Chassis, equipped with up to sixteen NVIDIA Tesla T4 GPUs contained in Amulet Hotkey CoreModules brings Tensor Core flexibility for data science and data analytic workloads.

Key factors:

  • NVIDIA Turing architecture with Tensor Cores to accelerate end-to-end data science workflows
  • Tesla T4 universal AI accelerator for data-driven visualization and machine learning
  • GPU-database solutions to leverage the massive parallel processing capability of GPUs including real-time processing to tap into the enormous value of insights from live data streams
  • Advanced SDKs and libraries including CUDA-X AI with RAPIDS as well as broad industry frameworks
  • NVIDIA Quadro vDWS software to virtualize workloads in VMware and Citrix environments
  • Bare-metal deployment using NVIDIA Tesla drivers and software

HPC Simulation and Modelling

Modern HPC data centers are key to solving some of the most important scientific and engineering challenges. HPC workloads require considerable compute and storage capacity to rapidly vast amounts of data to innovate and reduce the time to make new discoveries.

The Dell EMC PowerEdge MX7000 Modular Chassis, equipped with up to sixteen NVIDIA Tesla T4 GPUs contained in Amulet Hotkey CoreModules brings high-efficiency GPU acceleration for distributed computing environments running advanced research, simulation and modelling workloads.

Key factors:

  • High-efficiency scale-out performance that is ideal for distributed computing environments running single and mixed precision workloads
  • Delivers the latest NVIDIA Turing architecture and features including new Tensor Cores
  • Modern cloud paradigm using software-defined infrastructure flexibility to cover bare-metal, virtual and hybrid cloud workflows
  • NVIDIA Quadro vDWS software to virtualize applications in VMware and Citrix environments
  • Advanced SDK and libraries with broad industry framework and application support
  • GPU acceleration for over 600 HPC applications including ones infused with machine and deep learning technologies

Resources

Use GPU acceleration to support new and challenging workloads across the organization. Deploy in private, public or hybrid cloud environments that require performance, agility and flexibility. MX7000 servers with CoreModules combine Intel Xeon Scalable processor and NVIDIA Tesla GPU accelerator performance with the flexibility, scalability and efficiency of the MX7000 platform.

Documentation

CoreModule Datasheet

Core expansion module for PowerEdge MX7000

Download

Solutions Briefs

CoreModule GPU Solution Brief

Product and solution overview

Download

Find out which solution is right for you

Ask our Experts

Experts are ready to help configure a solution that meets your requirements

Contact

Additional Resources