Custom Hardware Acceleration in the AWS Cloud
The world’s fastest FPGA for accelerated computing is now accessible everywhere from the AWS Cloud.
Xilinx FPGAs are available in the Amazon Elastic Compute Cloud (Amazon EC2) F1 instances. F1 instances are designed to accelerate data center workloads including machine learning inference, data analytics, video processing, and genomics.
With F1 instances you can:
- Quickly deploy custom hardware acceleration
- Enjoy predictable performance with dedicated FPGAs
- Use your existing FPGA algorithms
in the AWS Cloud
Amazon EC2 F1 instances are available today in two different sizes that include up to eight Virtex® UltraScale+ VU9P FPGAs with a combined peak compute capability over 170 TOP/sec (INT8).
Xilinx Cloud Acceleration Tool on AWS
Applications and Services Running on F1 Instances
Amazon Web Services offers the Xilinx SDAccel ™ Development Environment for cloud acceleration. SDAccel empowers hardware designers to easily deploy their RTL designs in the F1 instance. It also automates the acceleration of code written in C, C++ or OpenCL by building application-specific accelerators on the F1 instances.
Get started with SDAccel in five easy steps.
If you prefer your dev system to be on premise, Vivado® Design Suite together with the SDAccel Development Environment can be purchased or upgraded to work with the F1 instances in the Amazon cloud. Purchase or upgrade now.
Academic Community
If you are interested in F1 instances as part of our academic partner program, go to the Xilinx University Program (XUP) cloud page.
Prerequisites
A number of preparation steps are required before being able to connect to an F1 instance. This module will show you how to:
- Create an AWS account
- Create an S3 bucket
- Preparate your private connection keys
- Request access to AWS F1 instances
1. Create, configure and test
This module will show you how to:
- Launch and configure an AWS F1 instance
- Connect to the instance and run a simple test
- Close the session
Follow the Instructions
2. Run your first SDAccel program
This module will show you how to:
- Package an existing RTL design as an SDAccel acceleration kernel
- Compile, execute and profile a simple example accelerated with the RTL kernel
Follow the Instructions
3. Learn about AWS F1 and SDAccel
This module will take you through an introductory online training course covering:
- Introduction to AWS F1 and the SDAccel Development Environment
- AWS F1 hardware specifications and requirements
- Packaging existing RTL IP for AWS F1
- Running a host-side application with an RTL kernel
4. Practice and experiment
This module will show you how to:
- Run additional labs and examples on F1
- Learn more about SDAccel
- Get support and troubleshoot issues
Follow the Instructions
5. Install and run SDAccel
The AWS F1 instances loaded with the FPGA developer AMI provide all the necessary tools to develop, compile, execute and debug your application in the Amazon cloud.
This optional module will show you how to:
- Install and license SDAccel locally on your machine
- Compile a simple example locally
- Upload and execute it on AWS F1
The following videos present key concepts needed to successfully run with existing RTL designs on AWS F1. We recommend watching these self-paced video modules in the order presented below.
Additional Resources
