industry news
Subscribe Now

NVIDIA Turing GPUs and NVIDIA Xavier Achieve Fastest Results on MLPerf Benchmarks Measuring Data Center and Edge AI Inference Performance

NVIDIA today posted the fastest results on new benchmarks measuring the performance of AI inference workloads in data centers and at the edge — building on the company’s equally strong position in recent benchmarks measuring AI training.

The results of the industry’s first independent suite of AI benchmarks for inference, called MLPerf Inference 0.5, demonstrate the performance of NVIDIA Turing™ GPUs for data centers and NVIDIA Xavier™ system-on-a-chip for edge computing.

MLPerf’s five inference benchmarks — applied across a range of form factors and four inferencing scenarios — cover such established AI applications as image classification, object detection and translation.

NVIDIA topped all five benchmarks for both data center-focused scenarios (server and offline), with Turing GPUs providing the highest performance per processor among commercially available entries1. Xavier provided the highest performance among commercially available edge and mobile SoCs under both edge-focused scenarios (single-stream and multi-stream)2.

“AI is at a tipping point as it moves swiftly from research to large-scale deployment for real applications,” said Ian Buck, general manager and vice president of Accelerated Computing at NVIDIA. “AI inference is a tremendous computational challenge. Combining the industry’s most advanced programmable accelerator, the CUDA-X suite of AI algorithms and our deep expertise in AI computing, NVIDIA can help data centers deploy their large and growing body of complex AI models.”

Watch a video of Buck discussing the MLPerf inference benchmarks: https://youtu.be/G3nsTSPY4LI

Highlighting the programmability and performance of its computing platform across diverse AI workloads, NVIDIA was the only AI platform company to submit results across all five MLPerf benchmarks. In July, NVIDIA won multiple MLPerf 0.6 benchmark results for AI training, setting eight records in training performance.

NVIDIA GPUs accelerate large-scale inference workloads in the world’s largest cloud infrastructures, including Alibaba Cloud, AWS, Google Cloud Platform, Microsoft Azure and Tencent. AI is now moving to the edge at the point of action and data creation. World-leading businesses and organizations, including Walmart and Procter & Gamble, are using NVIDIA’s EGX edge computing platform and AI inference capabilities to run sophisticated AI workloads at the edge.

All of NVIDIA’s MLPerf results were achieved using NVIDIA TensorRT™ 6 high-performance deep learning inference software that optimizes and deploys AI applications easily in production from the data center to the edge. New TensorRT optimizations are also available as open source in the GitHub repository.

New Jetson Xavier NX
Expanding its inference platform, NVIDIA today introduced Jetson Xavier NX, the world’s smallest, most powerful AI supercomputer for robotic and embedded computing devices at the edge. Jetson Xavier NX is built around a low-power version of the Xavier SoC used in the MLPerf Inference 0.5 benchmarks.

  1. MLPerf v0.5 Inference results retrieved from www.mlperf.org on Nov. 6, 2019, from entries Inf-0.5-15, Inf-0. 5-16, Inf-0.5-19, Inf-0.5-21. Inf-0.5-22, Inf-0.5-23, Inf-0.5-27. Per-processor performance is calculated by dividing the primary metric of total performance by number of accelerators reported.
  2. MLPerf v0.5 Inference results retrieved from www.mlperf.org on Nov. 6, 2019, from entries Inf-0.5-24, Inf-0.5-28, Inf-0.5-29.

About NVIDIA
NVIDIA’s (NASDAQ: NVDA) invention of the GPU in 1999 sparked the growth of the PC gaming market, redefined modern computer graphics and revolutionized parallel computing. More recently, GPU deep learning ignited modern AI — the next era of computing — with the GPU acting as the brain of computers, robots and self-driving cars that can perceive and understand the world. More information at http://nvidianews.nvidia.com/.

Leave a Reply

featured blogs
Jan 20, 2020
As you probably know, discrete wire component data is quite a bit different than just standard socket and terminal mating relationships. When we look at how Samtec approaches discrete wire products, there are several components involved. We not only sell the assemblies, but w...
Jan 20, 2020
My latest video blog is now available. This time I am looking at operating systems for embedded applications and how you go about selecting one. You can see the video here or here: Future video blogs will continue to look at topics of interest to embedded software developers....
Jan 17, 2020
I once met Steve Wozniak, or he once met me (it's hard to remember the nitty-gritty details)....
Jan 17, 2020
[From the last episode: We saw how virtual memory helps resolve the differences between where a compiler thinks things will go in memory and the real memories in a real system.] We'€™ve talked a lot about memory '€“ different kinds of memory, cache memory, heap memory, vi...

Featured Video

RedFit IDC SKEDD Connector

Sponsored by Wurth Electronics and Mouser Electronics

Why attach a header connector to your PCB when you really don’t need one? If you’re plugging a ribbon cable into your board, particularly for a limited-use function such as provisioning, diagnostics, or testing, it can be costly and clunky to add a header connector to your BOM, and introduce yet another component to pick and place. Wouldn’t it be great if you could plug directly into your board with no connector required on the PCB side? In this episode of Chalk Talk, Amelia Dalton chats with Ben Arden from Wurth Electronics about Redfit, a slick new connector solution that plugs directly into standard via holes on your PCB.

Click here for more information about Wurth Electronics REDFIT IDC SKEDD Connector