Vollo SDK

The Vollo SDK is designed for low latency streaming inference of machine learning (ML) models on FPGA platforms.

Installation

The latest SDK is available for download from https://github.com/MyrtleSoftware/vollo-sdk/releases.

Download the vollo-sdk-<version>.run self-extractable archive and execute it to extract the Vollo SDK contents to the current directory:

chmod +x vollo-sdk-<version>.run
./vollo-sdk-<version>.run

User guide

The user guide is available online at https://vollo.myrtle.ai/.

It is also included locally as markdown at docs/mdbook/src or in the release as html:

open docs/html/index.html

Overview

The product architecture is shown in the diagram below:

Vollo provides a C API to the user, running on the system host CPU. Vollo targets FPGAs to provide low latency inference. The FPGA images (bitstreams) are prebuilt and included in the product.

Vollo consists of the following elements:

• Vollo Accelerator Bitstream. Programming file for the FPGA on the PCIe accelerator card. See the Accelerator Setup section for instructions on programming your FPGA with the bitstream.

• Vollo Compiler. Compiles ML models defined in PyTorch or ONNX to Vollo programs.

• Vollo Runtime. The runtime library for Vollo. It sets up the the Vollo accelerator with a program and provides an asynchronous inference interface for handling input and output for the accelerated model.

Vollo Compiler

The Vollo compiler is available to use without an accelerator card or license. Along with compiling machine learning models for the Vollo accelerator, it also provides a simulation of the accelerator which can be used to provide a performance estimate.

Refer to the Vollo Compiler section in the user guide and its example walkthroughs to get started with the compiler.

Release file structure

Directory	Contents
bin/	Prebuilt applications (vollo-tool, vollo-onnx)
bitstream/	FPGA programming files
docs/	Documentation
example/	Example application and benchmark script
include/	Vollo runtime C/C++ header files
kernel_driver/	Kernel driver for Vollo accelerator card
lib/	Vollo runtime shared/static library files
python/	Vollo compiler Python libraries

Contact

For support and feature requests, please contact [email protected].