• Stars
    star
    105
  • Rank 321,612 (Top 7 %)
  • Language Tcl
  • Created over 4 years ago
  • Updated 5 months ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Vitis Embedded Platform Source Repository

Welcome to the Vitis embedded platform source repository. This repository contains the source code needed to recreate, modify, and extend the Xilinx-provided Vitis embedded platforms.

To build the Vitis embedded platforms from source code in this repository, you will need to have the following tools installed and follow the build instructions:

To learn how to customize Vitis embedded platforms, please refer to Vitis Platform Creation Tutorials.

This repository is divided into two sections:

  • The Xilinx_Official_Platforms directory contains official platform source provided by Xilinx. These official platforms are fully tested with all official Vitis examples, tutorials, etc.

  • The Xilinx_Reference_Platforms directory contains "unofficial" platform source. These platforms generally exist to illustrate particular concepts, provide specific I/O interfaces (e.g. MIPI and HDMI), etc. They are tested with the particular examples delivered for those platforms and Xilinx makes an effort to ensure compatibility, but they are not regressed and tested to the same degree as the official platforms.

Build Instructions

Vitis environment need to be setup before building the platform.

source <Vitis_install_path>/Vitis/2023.2/settings64.sh

This package comes with sources to generate the Vitis platform with these steps:

  1. Generate hardware specification file (XSA) using Vivado.
  2. Generate software components of platform (using XSCT).
  3. Generate the Vitis platform by packaging hardware and software together using XSCT tool

The following table summarizes use cases of platforms build :

Build Platform with pre-built Linux
Introduction: This method uses pre-built software components. Time taken for building sw components can be reduced
Requirements: Common Software. Use Sysroot SDK from Common Software
Pre-Synth command: make all PREBUILT_LINUX_PATH=<path/to/common_sw/dir>
Post-Impl command: make all PRE_SYNTH=FALSE PREBUILT_LINUX_PATH=<path/to/common_sw/dir>

The platform hardware has two types.

  • Pre-Synth XSA : Hardware specification file (XSA) in the platform does not contain bitstream. The XSA build time is quicker than Post-Impl XSA. By default, Vitis platform Makefile generates pre-synth platforms. Pre-Synth XSA is not valid for DFX platforms.
  • Post-Impl XSA : XSA generation flow goes through implementation. The XSA contains PL bitstream and the platform generation time will be longer.
    • DFX platforms require post-impl XSA.
    • Flat platform doesn't require post-impl XSA. If flat platform creation flow generates post-impl XSA, user can get platform resource info in platforminfo. Implementation issues may be found in early stage. It won't save time during the application v++ link phase. V++ will rerun implementation for the whole design.

Note: For Versal platforms, the PDI in XSA will always be pre-synth PDI.

There are two methods to prepare the platform software.

  • Use Pre-built Linux components: PetaLinux provides pre-built common software images. The image for each architecture includes Linux kernel, root file system, sysroot and common boot components like u-boot. The Vitis platform can use these components directly. Users need to download and install common images before building Vitis base platforms.
  • Use PetaLinux: Users can use Petalinux to generate everything locally. Using this method users can customize and generate software components. Users need to install PetaLinux and setup PetaLinux environment before building Vitis base platforms. Petalinux software components generation Tutorial

The Flags used in commands of above table are explained below :

  • PREBUILT_LINUX_PATH: By default, this flag is set to /opt/xilinx/platform/xilinx--common-. If platform needs to be build using package located at another directory then provide the pre-built Linux image path to this flag. The build flow will generate device tree from XSA and run platform package with the pre-built Linux image. Common Linux components are provided as pre-built binaries by Xilinx to fullfill most evaluation requirements. Please refer to UG1393- Using Embedded Platforms chapter for more information. Set this flag to point common software components(u-boot, boot.scr, bl31.elf). These files are architecture specific. Zynq, zynqMP and Versal have different software component files.

    make all
    # Platform will be build using images found at /opt/xilinx/platform/xilinx-<arch>-common-<ver> or in current dir
    
    make all PREBUILT_LINUX_PATH=<common_sw_dir> 
    # This dir should contain u-boot.elf, boot.scr and bl31.elf
  • PRE_SYNTH: This flag is used to specify XSA type that needs to be built. The default value is TRUE, which means pre-synth XSA is generated. To generate post-impl XSA, set this variable to FALSE. The build flow will run through implementation in Vivado and include implementation results to the post-impl XSA.

  • XSA: This flag is used when building a platform with a pre-defined XSA. Example usage:

    make linux XSA=/home/user/<platform_dir>/vivado/build/<platform_name>.xsa
    make platform

After the build process completes, the built platform output is located at platform_repo/<platform_name>/export/<platform_name>/ directory.

Notes:

  • Building of the base platform is supported on Linux environments only (although it is possible to build inside a VM or Docker container), because the platform building process involves cross-compiling in Linux.

Installing the Common Software

Xilinx provides pre-built Common Software images so that user can skip PetaLinux project creation and build process for common usages. They can be downloaded from Xilinx Download Center under Vitis Embedded Platform tab or PetaLinux tab. Please refer to Common images for Embedded Vitis platforms section.

The common image packages download contains these components

  • Prebuilt Linux kernel
  • Prebuilt root file system
  • sdk.sh to generate sysroot
  • boot files (pre-built u-boot.elf, boot.scr, bl31.elf, etc)

You can use the following command to extract the tar image.

tar -xzf xilinx-<arch>-common-<version>.tar.gz

To use the common software, download pre-built common images from https://www.xilinx.com/support/download/index.html/content/xilinx/en/downloadNav/embedded-platforms.html and extract the common images to /opt/xilinx/platform/ or pass it as PREBUILT_LINUX_PATH parameter make command.

Installing the Sysroot (Yocto SDK)

Vitis applications require to build with libaries in sysroot (Yocto SDK). Sysroot can be installed from sdk.sh, which is included in Common Software images. Building Vitis base platforms doesn't require sysroot.

To install sysroot from sdk.sh, you can run the script sdk.sh directly.

./sdk.sh

You can also check the available options for scripting.

./sdk.sh --help
Usage: sdk.sh [-y] [-d <dir>]
  -y         Automatic yes to all prompts
  -d <dir>   Install the SDK to <dir>
======== Extensible SDK only options ============
  -n         Do not prepare the build system
  -p         Publish mode (implies -n)
======== Advanced DEBUGGING ONLY OPTIONS ========
  -S         Save relocation scripts
  -R         Do not relocate executables
  -D         use set -x to see what is going on
  -l         list files that will be extracted

To cross-compile with this SDK from the command line, run source ./environment-setup-*-xilinx-linux to set up your environment (cross-compile build tools, libraries, etc).

More Repositories

1

PYNQ

Python Productivity for ZYNQ
Jupyter Notebook
1,894
star
2

Vitis-AI

Vitis AI is Xilinx’s development stack for AI inference on Xilinx hardware platforms, including both edge devices and Alveo cards.
Python
1,400
star
3

linux-xlnx

The official Linux kernel from Xilinx
C
1,205
star
4

brevitas

Brevitas: neural network quantization in PyTorch
Python
1,107
star
5

Vitis-Tutorials

Vitis In-Depth Tutorials
C
891
star
6

Vitis_Libraries

Vitis Libraries
C++
818
star
7

embeddedsw

Xilinx Embedded Software (embeddedsw) Development
HTML
766
star
8

finn

Dataflow compiler for QNN inference on FPGAs
Python
679
star
9

BNN-PYNQ

Quantized Neural Networks (QNNs) on PYNQ
Jupyter Notebook
652
star
10

u-boot-xlnx

The official Xilinx u-boot repository
C
531
star
11

XRT

Run Time for AIE and FPGA based platforms
C++
529
star
12

Vitis_Accel_Examples

Vitis_Accel_Examples
Makefile
467
star
13

Vitis-HLS-Introductory-Examples

C++
420
star
14

dma_ip_drivers

Xilinx QDMA IP Drivers
C
400
star
15

HLS

Vitis HLS LLVM source code and examples
375
star
16

Vitis-AI-Tutorials

354
star
17

PYNQ_Workshop

Jupyter Notebook
354
star
18

SDAccel_Examples

SDAccel Examples
C++
350
star
19

ml-suite

Getting Started with Xilinx ML Suite
Jupyter Notebook
334
star
20

CHaiDNN

HLS based Deep Neural Network Accelerator Library for Xilinx Ultrascale+ MPSoCs
C++
315
star
21

xfopencv

C++
313
star
22

XilinxTclStore

Xilinx Tcl Store
Tcl
310
star
23

mlir-aie

An MLIR-based toolchain for AMD AI Engine-enabled devices.
MLIR
257
star
24

RapidWright

Build Customized FPGA Implementations for Vivado
Java
248
star
25

QNN-MO-PYNQ

Jupyter Notebook
234
star
26

XilinxBoardStore

Python
224
star
27

libsystemctlm-soc

SystemC/TLM-2.0 Co-simulation framework
Verilog
200
star
28

qemu

Xilinx's fork of Quick EMUlator (QEMU) with improved support and modelling for the Xilinx platforms.
C
200
star
29

DPU-PYNQ

DPU on PYNQ
Tcl
189
star
30

device-tree-xlnx

Linux device tree generator for the Xilinx SDK (Vivado > 2014.1)
Tcl
181
star
31

PYNQ-ComputerVision

Computer Vision Overlays on Pynq
Jupyter Notebook
173
star
32

XilinxVirtualCable

Xilinx Virtual Cable (XVC) is a TCP/IP-based protocol that acts like a JTAG cable and provides a means to access and debug your FPGA or SoC design without using a physical cable.
C
172
star
33

finn-hlslib

Vitis HLS Library for FINN
C++
168
star
34

graffitist

Graph Transforms to Quantize and Retrain Deep Neural Nets in TensorFlow
Python
168
star
35

open-nic

AMD OpenNIC Project Overview
Shell
166
star
36

finn-examples

Dataflow QNN inference accelerator examples on FPGAs
Jupyter Notebook
163
star
37

SDSoC-Tutorials

SDSoCâ„¢ (Software-Defined System-On-Chip) Environment Tutorials
C++
142
star
38

xilinx-tiny-cnn

C++
140
star
39

FPGA_as_a_Service

Go
136
star
40

xup_vitis_network_example

VNx: Vitis Network Examples
Jupyter Notebook
124
star
41

meta-xilinx

Collection of Yocto Project layers to enable AMD Xilinx products
C
123
star
42

Vitis-In-Depth-Tutorial

C++
113
star
43

systemctlm-cosim-demo

QEMU libsystemctlm-soc co-simulation demos.
C++
106
star
44

SDAccel-Tutorials

SDAccel Development Environment Tutorials
C++
101
star
45

nanotube

LLVM
101
star
46

RFNoC-HLS-NeuralNet

CMake
92
star
47

Embedded-Design-Tutorials

91
star
48

PYNQ-DL

Xilinx Deep Learning IP
VHDL
91
star
49

PYNQ-HelloWorld

This repository contains a "Hello World" introduction application to the Xilinx PYNQ framework.
Jupyter Notebook
90
star
50

LSTM-PYNQ

C++
86
star
51

Vivado-Design-Tutorials

Tcl
83
star
52

SDSoC_Examples

C++
82
star
53

Kria-PYNQ

PYNQ support and examples for Kria SOMs
Jupyter Notebook
82
star
54

meta-petalinux

meta-petalinux distro layer supporting Xilinx Tools
BitBake
82
star
55

kria-vitis-platforms

Kria KV260 Vitis platforms and overlays
SystemVerilog
81
star
56

IIoT-EDDP

The repository contains the design database and documentation for Electric Drives Demonstration Platform
VHDL
79
star
57

logicnets

Python
78
star
58

AI-Model-Zoo

75
star
59

RecoNIC

RecoNIC is a software/hardware shell used to enable network-attached processing within an RDMA-featured SmartNIC for scale-out computing.
SystemVerilog
75
star
60

ACCL

Alveo Collective Communication Library: MPI-like communication operations for Xilinx Alveo accelerators
C++
75
star
61

open-nic-shell

AMD OpenNIC Shell includes the HDL source files
SystemVerilog
70
star
62

mlir-air

C++
70
star
63

Applications

C
67
star
64

llvm-aie

Fork of LLVM to support AMD AIEngine processors
LLVM
66
star
65

XilinxUnisimLibrary

Xilinx Unisim Library in Verilog
Verilog
64
star
66

PYNQ_Composable_Pipeline

PYNQ Composabe Overlays
Tcl
61
star
67

gemx

Matrix Operation Library for FPGA https://xilinx.github.io/gemx/
C++
56
star
68

PYNQ_RFSOC_Workshop

Open-sourcing the PYNQ & RFSoC workshop materials
Jupyter Notebook
55
star
69

merlin-compiler

C++
52
star
70

meta-xilinx-tools

Yocto Project layer enables AMD Xilinx tools related metadata for MicroBlaze, Zynq, ZynqMP and Versal devices.
BitBake
50
star
71

RFSoC-PYNQ

Python productivity for RFSoC platforms
Jupyter Notebook
49
star
72

ResNet50-PYNQ

Quantized ResNet50 Dataflow Acceleration on Alveo, with PYNQ
C++
48
star
73

Alveo-PYNQ

Introductory examples for using PYNQ with Alveo
Jupyter Notebook
47
star
74

xup_compute_acceleration

Hands-on experience using the Vitis unified software platform with Xilinx FPGA hardware
C++
46
star
75

xup_high_level_synthesis_design_flow

AMD Xilinx University Program HLS tutorial
C
46
star
76

Vitis_Model_Composer

Vitis Model Composer Examples and Tutorials
C++
46
star
77

Vitis-AWS-F1-Developer-Labs

C++
44
star
78

PYNQ_Bootcamp

PYNQ Bootcamp 2019-2022 teaching materials.
Jupyter Notebook
44
star
79

PYNQ-Networking

Networking Overlay on PYNQ
Tcl
44
star
80

KRS

The Kria Robotics Stack (KRS) is a ROS 2 superset for industry, an integrated set of robot libraries and utilities to accelerate the development, maintenance and commercialization of industrial-grade robotic solutions while using adaptive computing.
HTML
43
star
81

Get_Moving_With_Alveo

For publishing the source for UG1352 "Get Moving with Alveo"
C++
42
star
82

blockchainacceleration

Tcl
42
star
83

HLS_packet_processing

C++
41
star
84

HLS_arbitrary_Precision_Types

C++
40
star
85

DSRL

40
star
86

inference-server

C++
40
star
87

Xilinx_Kria_KV260_Workshop

39
star
88

chipscopy

ChipScoPy (ChipScope Python API) is an open source Python API to the various ChipScope services provided by the TCF-based (Target Communication Framework) ChipScope Server (cs_server).
Jupyter Notebook
38
star
89

VVAS

Vitis Video Analytics SDK
C
37
star
90

vcu-ctrl-sw

C
36
star
91

XilinxCEDStore

This store contains Configurable Example Designs.
Tcl
36
star
92

pyxir

Python
36
star
93

pytorch-ocr

Python
35
star
94

DSP-PYNQ

A PYNQ overlay demonstrating Pythonic DSP running on Zynq UltraScale+
Tcl
35
star
95

xup_aie_training

Hands-on experience programming AI Engines using Vitis Unified Software Platform
Jupyter Notebook
34
star
96

open-nic-driver

AMD OpenNIC driver includes the Linux kernel driver
C
33
star
97

pcie-model

PCI Express controller model
C
32
star
98

qemu-devicetrees

Device trees used by QEMU to describe the hardware
Makefile
32
star
99

bootgen

bootgen source code
C++
31
star
100

hdmi-modules

Xilinx Soft-IP HDMI Rx/Tx core Linux drivers
C
30
star