We have mostly discussed the DPU for use with edge AI applications. If you are wondering if the DPU can also be used for Data Center or on-premise AI applications, the answer is yes!
We recently demonstrated the DPU on an Alveo data center acceleration card at the CVPR 2019 conference.
We showed a Resnet50 demo which runs on the Alveo U250 data center accelerator card with an industry-leading mixed precision implementation. This demo used Int8/Int2 activation and Int8/Ternary weights.
With the DPU design optimized for the Alveo U250 data center accelerator card, it can run Resnet50 @ 5100+ fps and around 3ms latency with batch size of 16. By comparison, with batch size of 32, Nvidia T4 runs Resnet50 @ 4600+ fps (6.8ms latency!)
The demo screenshot is shown below.
More from AI and Machine Learning Blog articles
ADAS Front Camera with Xilinx-powered AI has entered mass production at BYD
Xilinx Zynq Devices Take Charge of Robots
Xilinx at the Global AI Product Application Expo 2019
Uncanny AI Box for ALPR Application with Kria KV260 Starter Kit
Xilinx Champions AI at ISE 2020
From Xilinx Challenge Winner to App Store Seller!
![](https://platodata.io/wp-content/uploads/2021/07/study-bitcoin-and-dogecoin-are-the-first-investment-for-45-of-young-britons-300x180.jpg)
Xilinx and ZF to Jointly Enable AI Innovation for Autonomous Driving
AI @ Xilinx Technology Day – Shanghai
![](https://platodata.io/wp-content/uploads/2021/05/another-btc-panic-sell-off-looming-blockchain-analytics-firm-glassnode-weighs-in-300x169.png)
Low-latency HD Inference – a New Treatment for Myopic Vision Systems
Single-Shot Detection (SSD) Tutorials
![](https://platodata.io/wp-content/uploads/2021/05/uk-mp-pressures-rishi-sunak-over-delays-to-fca-crypto-register-300x169.jpg)