Unleash Data Center Performance with Xilinx® Alveo™ Accelerator Solutions
Xilinx® Alveo™ Accelerator Powered Workstations and Servers from Exxact are designed to meet the constantly changing needs ofthe modern data center, providing up to 90X performance increase over CPUs for computationally intensive workloads.
HIGH PERFORMANCE WORKLOAD
Our Xilinx® Alveo™ powered workstations and servers perform up to 90X higher than CPUs on key workloads at 1/3 the cost.
ADAPTABLE ACCELERATION
Exxact systems with Xilinx® Alveo™ Accelerators can perform multiple tasks from machine learning inference, to video processing, to any workload using the same accelerator card.
RECONFIGURABLE HARDWARE
As workload algorithms evolve, reconfigurable hardware such as Xilinx® Alveo™ Accelerators adapt faster than fixed-function accelerator card product cycles.
FLEXIBILITY FOR DEPLOYMENT
Deploy FPGA accelerated solutions in the cloud or on-premises interchangeably, scalable to application requirements.
PREINSTALLED XILINX ML SUITE
Anaconda dev environment for implementing ML software (TensorFlow, Keras, MXNET, Caffe, etc) to run on Xilinx® Alveo™ Accelerators.
PYTHON BASED QUANTIZER
Take existing pre-trained models and optimize them for the Xilinx® Alveo™ Accelerators, with no re-training needed.
Suggested XILINX® ALVEO™ ACCELERATOR Systems
Entry-Level
Valence Workstation
2x 1TB HDD (Data)
INQUIRE FOR PRICING
Mid-Range
Valence Workstation
2x 2TB HDD (Data)
INQUIRE FOR PRICING
High-End
Tensor Server
2x 2TB HDD (Data)
CALL FOR PRICING
Not Sure What You Need?
Let us know what kind of project you have
planned and we can help you decide.
TELL US WHAT YOU WANT TO DO
Xilinx® Alveo™ Accelerated System Use Cases
Database Search
and Analytics
Financial
Computing
Machine
Learning
Video Processing
and Transcoding
Genomic
Sequencing
Solve Problems Now with Xilinx ML Suite Preinstalled
Step by Step Tutorials to
Run Custom Network Models
JUPYTER NOTEBOOK TUTORIALS
xfDNN Tools
- Using the xfDNN Compiler with Caffe models
- Using the xfDNN Quantizer to quantize Caffe models
Image Classification
Object Detection
» FIND THEM ON GITHUB
COMMAND LINE TUTORIALS
xfDNN Tools- Compiling networks with xfDNN Compiler
- Calibrate high precision models for INT16/8 Deployment
Object Detection
Web Services
- DeepDetect REST Tutorial
- DeepDetect Webcam Tutorial
Image Classification
- Image classification with GoogLeNet v1 Demo on VCU1525
- Image classification with Python APIs
Xilinx® Alveo™ Accelerated Systems For
Advanced Video Processing & Live Streaming Applications
High definition video, images, streaming and video transcoding require accelerated performance to provide a consistent user viewing experience.
Alveo Advantage
- 12x increase in encoding performance
- 10x lower power
- Half the cost per channel
- Integrates into existing infrastructure via FFmpeg
Specifications
Workstation
| Peak INT8 TOPs | 18.6 |
| Mem. Bandwidth | 77GB/s DDR |
| Intl. SRAM Bandwidth | 31TB/s |
| Look Up Tables | 892,000 |
| Cooling | Active |
| LEARN MORE |
Server
| Peak INT8 TOPs | 18.6 |
| Mem. Bandwidth | 77GB/s DDR |
| Intl. SRAM Bandwidth | 31TB/s |
| Look Up Tables | 892,000 |
| Cooling | Passive |
| LEARN MORE |
Workstation
| Peak INT8 TOPs | 33.3 |
| Mem. Bandwidth | 77GB/s DDR |
| Intl. SRAM Bandwidth | 38TB/s |
| Look Up Tables | 1,341,000 |
| Cooling | Active |
| LEARN MORE |
Server
| Peak INT8 TOPs | 33.3 |
| Mem. Bandwidth | 77GB/s DDR |
| Intl. SRAM Bandwidth | 38TB/s |
| Look Up Tables | 1,341,000 |
| Cooling | Passive |
| LEARN MORE |
Xilinx® Alveo™ Accelerated Systems
Alveo Data Center accelerator cards can deliver dramatic acceleration across a broad set of applications and are reconfigurable to provide an ideal fit for the changing workloads of the modern data center.
With workloads evolving faster than silicon design cycles, Xilinx FPGAs can keep pace. Built on Xilinx 16nm UltraScale architecture, Alveo U200 and U250 accelerator cards provide reconfigurable acceleration that can adapt to continual algorithm optimizations, supporting multiple workload types while reducing overall cost of ownership.
FILTER
