Essential DSP Implementation Techniques for Xilinx FPGAs

SO_DSP_BASIC

Course Description

This course provides a foundation for Digital Signal Processing (DSP) techniques for AMD Xilinx FPGAs. The course begins with a refresher of basic binary number theory, mathematics, and the essential features within the FPGA that are important to signal processing. The body of the course explores a variety of filter techniques with emphasis on optimal implementation in AMD Xilinx devices and continues with an examination of FFTs, video, and image processing. Throughout the course, AMD Xilinx cores and IP relevant to signal processing are introduced. The course is complemented by hands-on exercises to reinforce the concepts learned.

Release Date

March 2015

Level

DSP 3

Training Duration

2 days

Who Should Attend?

Engineers and designers who have an interest in developing products that use digital signal processing.

Prerequisites

A fundamental understanding of digital signal processing theory, including an understanding of the following principles

• Sample rates
• Finite Impulse Response (FIR) and Infinite Impulse Response (IIR) filters
• Oscillators and mixers
• Fast Fourier Transform (FFT) algorithm

Recommended

Essentials of FPGA Design course

Skills Gained

After completing this training, you will know how to:

• Describe the advantages of using FPGAs over traditional processors for DSP designs
• Utilize fixed point binary arithmetic and identify how to use this knowledge to create efficient designs in FPGAs
• Recognize how both the CLB slices in FPGAs and the more advanced DSP48s are used to implement DSP algorithms
• Explain the dataflow through the device and how to use distributed memory, block RAM, registers, and SRLs to properly implement these designs
• Construct different FIR filter and FFT implementations and how to optimize these implementations in the FPGA
• Explain the algorithms for video and imaging systems and their implementations in FPGAs

Course Outline

Day 1

• Back to Basics
• Architecture
• FPGA Math
• Exercise 1: Signed Number Conversion, Quantization and Rounding, Adders, Subtractors, and Accumulation
• Shift Registers, RAM, and Applications
• Exercise 2: SRL32E and RAM Estimation and Concatenation
• FIR Filter
• Exercise 3: Filter Implementation, Resource and Performance Estimation

Day 2

• Advanced Filter Techniques
• Exercise 4: Filter Implementations, Resource and Performance Estimation
• Fast Fourier Transform
• Exercise 5: FFT Implementation, Resource and Performance Estimation
• Video and Imaging
• Where Do We Go From Here?
• Demonstration: System Generator and the CORE Generator Tool with a DSP-Targeted Reference Design
• Where Can I Learn More?

Course Exercises

• Exercise 1: Signed Number Conversion, Quantization and Rounding, Adders, Subtractors, and Accumulation – Learn how to estimate device resource utilization for basic math functions. Compare different methodologies for implementing functions. 
• Exercise 2: SRL32E and RAM Estimation and Concatenation – Learn how to optimize memory and storage in AMD Xilinx FPGAs.
• Exercise 3: Filter Implementation, Resource and Performance Estimation – Learn how and when to use various implementation strategies for optimal filter implementation. 
• Exercise 4: Filter Implementations, Resource and Performance Estimation – Advanced filter topologies are studied. Architect multichannel and multirate filters using various methods.  Implementation strategies will be discussed and optimal methods used.
• Exercise 5: FFT Implementation, Resource and Performance Estimation – Select correct parameters for FFT implementations to meet design targets. Resource estimation will be studied and trade-offs with performance examined through implementation examples.
• Demonstration: System Generator and the CORE Generator Tool with a DSP-Targeted Reference Design – Introduces DSP-targeted hardware boards and software tools. Witness the power, ease of use, and design efficiency of AMD Xilinx DSP tools and IP.  Reinforce the concepts studied in the course material and exercises.