The Vitis unified IDE and v++ common
command line support foundational C/C++ libraries allowing common hardware design
constructs and functions to be easily modeled in C/C++ and synthesized to RTL. The
following libraries are provided with with the tools as described in Vitis HLS Libraries Reference:
- Arbitrary Precision Data Types Library: Arbitrary precision data types let your C code use variables with smaller bit-widths than standard C or C++ data types, to enable improved performance and reduced area in hardware.
- HLS Math Library: Used to specify standard math operations for synthesis into RTL and implementation on AMD devices.
- HLS Stream Library: For modeling and compiling streaming data structures.
You can use each of the C libraries in your design by including the
library header file in your code. These header files are located in the include directory in the Vitis HLS installation area.
Vitis Libraries
In addition, the Vitis
Libraries are available for use with Vitis unified
IDE and v++ common command line. The libraries
include common functions of math, statistics, linear algebra and DSP; and also
supporting domain specific applications, like vision and image processing,
quantitative finance, database, data analytics, and data compression. The libraries
can be downloaded from https://github.com/Xilinx/Vitis_Libraries and setup for use in the Libraries view in the Vitis unified IDE, as described in
Working with the Vitis
Unified IDE in the Vitis Unified
Software Platform Documentation: Application Acceleration
Development (UG1393).
The libraries contain functions and constructs that are optimized for implementation on AMD devices. Using these libraries helps to ensure high quality of results (QoR); that the results of synthesis are a high-performance design that optimizes resource usage. Because the libraries are provided in C and C++, you can incorporate the libraries into your top-level function and simulate them to verify the functional correctness before synthesis.
Working with OpenCV
As described at https://github.com/Xilinx/Vitis_Libraries/tree/main/vision, to work with the Vitis Vision Library you must also have installed the OpenCV computer vision library. Vitis HLS does not come with OpenCV pre-installed, but the following link describes how to find and install it: Vitis Libraries - Compiling and Installing OpenCV libraries for use with Vision library.