insideHPC Guide to NVIDIA EGX Platform: One Architecture for Every Workload – Part 3 – insideHPC – insideHPC

Sponsored Post

Introduction

The massive growth of data requires that organizations analyze their data center infrastructure to meet changing workload needs. Artificial Intelligence (AI), High Performance Computing (HPC), big data analytics and professional visualization are growth drivers which are changing workloads. AI is foundational and is  transforming the way businesses work across many industries. But the use of AI will become critical as new  tools and technologies such as 5G wireless hyperconnectivity, edge computing, and Radio Area Networks  (RANs) become a reality. According to Carl Flygare, NVIDIA Pro GPU Product Marketing Manager, PNY  Technologies, “Organizations need to reimagine data centers to meet AI and big data analytics intensive  workload needs, sophisticated visualization and rendering workflows, and prepare for a GPU accelerated VDI enabled mobile workforce.”

This technology guide, “insideHPC Guide to NVIDIA EGX Platform: One Architecture for Every Workload,” sponsored by PNY describes how the NVIDIA® EGX™ Platform provides a single architecture to meet every data center workload need. PNY aids in GPU selection and testing allowing partners to equip customers with the  best NVIDIA EGX Platform solution based on their needs.

NVIDIA Software Meets Customer Needs

The NVIDIA EGX Platform solution contains far more than hardware. It also provides a comprehensive integrated software solution including container orchestration, management integration, and a variety of  virtualization tools as well as third-party hypervisors from all major vendors and supports Windows or Linux operating systems. The platform also encompasses an ecosystem that accelerates applications and frameworks developed and delivered by NVIDIA NGC™, which features over 100 pre-trained AI models and other applications, or leading Independent Software Vendors (ISVs).

There are over 100 applications that are fully optimized to run on NVIDIA RTX GPUs. These solutions are in areas such as AI training and inference, edge AI, data analytics, machine learning, professional visualization, and collaboration. All NVIDIA software initiatives can run on the NVIDIA EGX Platform in bare metal or virtualized deployments. NVIDIA provides enterprise class support for this mission critical software.

NVIDIA Virtual PC (vPC) and Virtual Apps (vApps) improve virtual desktops and applications. The virtualization software divides NVIDIA GPU resources, so the GPU can be shared across multiple virtual machines running any application.

NVIDIA RTX Virtual Workstation (vWS) supports advanced features of NVIDIA RTX for graphics, visualization,  and AI. Using NVIDIA RTX technology, vWS brings advanced features like ray tracing, AI-denoising, and Deep  Learning Super Sampling (DLSS) to a virtual environment. The application can virtualize any application from  the data center with an experience that is indistinguishable from a physical workstation—enabling  workstation performance from any device.

NVIDIA AI Enterprise (NVAIE) is an end-to-end, cloud native software suite of AI and data science applications  and frameworks optimized and exclusively certified by NVIDIA to run on VMware vSphere with  NVIDIA-Certified Systems based on the NVIDIA EGX Platform. It includes key enabling technologies and best- in-class AI software from NVIDIA for rapid deployment, management and scaling of AI workloads in the  modern hybrid cloud.

PNY’s Role in the NVIDIA EGX Platform Ecosystem

Established in 1985, PNY is a global technology leader dedicated to business-grade and consumer electronics  manufacturing with over 30 years of business experience serving consumers, B2Bs, and OEMs  worldwide. The company works with its large and varied pool of system builders to offer NVIDIA EGX Platform-based solutions across all major use cases and deployment scenarios. When designing a system, PNY consults with customers and select system builder partners to determine workload needs and then  jointly recommends the right GPU for business workloads.

NVIDIA Certified Systems are thoroughly and rigorously tested with specialized NVIDIA software and real- world workloads that validate the best system performance and stability GPUs, CPUs, Smart NICs for Smart  NICs, mass storage, system memory, PCIe controller performance, and system enclosure thermal management. Once a system passes these tests it can be marketed and sold as an NVIDIA-Certified System, which will deliver the highest levels of performance and reliability to the purchaser.

NVIDIA EGX Platform Benefits

  • Security: Data is secured and accessed from the data center with appropriate permissions which lessens the chance of critical intellectual property (IP) being compromised
  • Accelerating Time to Market while reducing costs
  • Manageability: Scale resources to easily add new virtual workstation users in minutes
  • Collaborate from anywhere: Accelerate creative process by giving employees freedom to access fully 3D capable workstation on any device from anywhere

NVIDIA EGX Platform Power, Performance, and Cost Benefits

The NVIDIA EGX Platform and NVIDIA GPUs provide exponential power at a fraction of the cost of CPU based alternatives, which reduces acquisition costs and dramatically reduces the total cost of ownership. NVIDIA  GPUs provide increased FP32, F64 AI, HPC, and rendering performance. Many optimizations are provided for  each. For example, the NVIDIA Ampere-based GPU micro architecture supports structural sparsity which can  double the performance of non-sparse models. Figure 4 shows the increased processing performance of the  NVIDIA EGX Platform solution compared to CPUs.

Lost in Space Example

An experiment recreated scenes from the Netflix Lost in Space series using RTX GPUs to compare the results with the initial CPU-based infrastructure. In terms of performance, the solution had six times faster processing for a single frame with a 25-fold rendering advantage. In addition, there was a seven-fold cost improvement including acquisition cost, five-year cost and cost of power.

Summary

Organizations are increasingly using AI, HPC, big data analysis, and professional visualization which is driving  data growth and changing data center workloads. Many data centers have legacy systems which rely on CPUs  for such processing. Since CPUs essentially process data serially, even with multi-core designs and  techniques such as hyperthreading, CPUs inherently slower performance than massively parallel GPUs cannot handle currently evolving data center workload needs. “The growing use of AI will be a requirement for  organizations to remain competitive. Everyone must evaluate their future data center needs to meet AI and  big data analytics intensive workload challenges and opportunities and prepare for a mobile workforce,” states Flygare.

NVIDIA GPUs and the NVIDIA EGX Platform are designed to meet the workflow needs of modern data  centers on a single platform. The NVIDIA solution meets data processing needs for HPC, AI, DL, big data  analytics as well as professional visualization processing and rendering needs. PNY is a global technology  leader who can support organizations in the selection, testing, and implementation of NVIDIA GPU solutions and the NVIDIA EGX Platform.

Future Role of NVIDIA Solutions

NVIDIA is working to expand the role GPUs and other NVIDIA products and technologies will play in Exascale data centers. At the GPU Technology Conference (GTC 2021), NVIDIA CEO Jensen Huang described some  world-changing technologies based on NVIDIA GPUs and solutions such as the NVIDIA EGX Platform. One  solution is the Earth-2 (E2) supercomputer that NVIDIA plans to develop as a tool to help mitigate climate change effects.

NVIDIA also announced that the company is collaborating with Google Quantum AI, IBM, and others to take  quantum computing to the next level. Quantum computers, when fully realized, can solve problems in fields  like drug development, climate research, machine learning and finance. The potential is great but there are  many challenges. NVIDIA is developing cuQuantum software, libraries and installing this software on select  NVIDIA data center GPU boards to aid in the development of practical quantum computers.

“These moonshot technological advancements, including advancing the development of quantum computing, announced by NVIDIA have the ability to fundamentally change our understanding of the physical and natural sciences, large swaths of life as we know it, and even change our understanding of physical reality itself” states Flygare.

Visit www.pny.com/egx-platform and complete the online form to be contacted by one of our partners for additional information on NVIDIA EGX, or to begin the purchase process of an NVIDIA EGX compliant system (or multiple systems).

Over the past few weeks we’ve explored these topics:

Download the complete insideHPC Guide to NVIDIA EGX Platform: One Architecture for Every Workload courtesy of PNY

Spread the love

Leave a Reply

Your email address will not be published.