Máy chủ trí tuệ nhân tạo NVidia DGX-1

The Challenges of Building a Platform for AI

Data scientists depend on computing performance to gain insights and innovatefaster, using the power of deep learning and analytics. GPU technology offers afaster path to AI, but building a platform goes well beyond deploying a serverand GPU’s.AI and deep learning can require a substantial commitment in softwareengineering. An investment that could delay your project by months as youintegrate a complex stack of components and software includingframeworks, libraries, and drivers. Once deployed, additional time andresources are continually needed as you wait for the ever-evolving opensource software to stabilize. You’ll also be waiting to optimize yourinfrastructure for performance, along with administrative costs thatincrease as the system scales.

The Fastest Path to Deep Learning

Inspired by the demands of AI and data science, NVIDIA® DGX-1™ fast-tracksyour AI initiative with a solution that works right out of the box so that youcan gain insights in hours instead of months. With DGX-1 you can simply plugin, power up, and get to work, thanks to the integrated NVIDIA deep learningsoftware stack. Now you can start deep learning training in as little as a day,instead of spending months integrating, configuring, and troubleshootinghardware and software.

Effortless Productivity

NVIDIA DGX-1 removes the burden of continually optimizing your deep learningsoftware and delivers a ready-to-use, optimized software stack that can saveyou hundreds of thousands of dollars. It includes access to today’s most populardeep learning frameworks, NVIDIA DIGITS™ deep learning training application,third-party accelerated solutions, the NVIDIA Deep Learning SDK (e.g. cuDNN,cuBLAS, NCCL), CUDA® toolkit, Docker Engine Utility for NVIDIA GPU.

Revolutionary AI Performance

While many solutions offer GPU-acceleratedperformance, only NVIDIA DGX-1 unlocks thefull potential of the latest NVIDIA GPU’s like theNVIDIA Tesla® V100, including innovations likenext generation NVLink™ and new Tensor Corearchitecture. With its performance-engineereddeep learning software stack, DGX-1 delivers upto three times faster training speed than otherGPU-based systems. With the computing capacityof 140 servers in a single system that integrates thelatest NVIDIA GPU technology with the world’s mostadvanced deep learning software stack, you cantake advantage of revolutionary performance to gaininsights faster than ever, powered by NVIDIA DGX-1.

Investment Protection

Your AI initiative is critical to your organization’ssuccess, and dependent on a frequently optimizedsoftware stack and integrated hardwareinfrastructure. With today’s rapidly evolving opensource software and the complexity of libraries,drivers, and hardware, it’s good to know thatNVIDIA’s enterprise grade support and softwareengineering expertise are behind every DGX-1. Thissoftware stack is built on years of R&D, innovation,and deep learning expertise, and maintained bymonthly optimized framework releases. Also,NVIDIA's support includes software upgrades andpriority resolution of critical issues; you can havepeace of mind that your environment is tuned formaximized performance and uptime.