Introduction to FPGA Design Using MATLAB and Simulink

Eric Cigan MathWorks

Jordon Inkeles Altera

2011 The MathWorks, Inc. 1

Separate Views of FPGA Design

System Designer FPGA Designer

Algorithm Design
Fixed-Point Timing and Control Logic Architecture Exploration Algorithms / IP

System Test Bench

Environment Models Analog Models Digital Models Algorithms / IP

RTL Design
IP Interfaces Hardware Architecture

Verification
Behavioral Simulation Functional Simulation Static Timing Analysis Timing Simulation

Implement Design FPGA Requirements

Hardware Specification

Back Annotation

Synthesis
Map

Place & Route

Test Stimulus

FPGA Hardware

Model-Based Design for FPGAs

MATLAB and Simulink System and Algorithm Design
Algorithm Design
Fixed-Point Timing and Control Logic Architecture Exploration Algorithms / IP

System Test Bench

Environment Models Analog Models Digital Models Algorithms / IP

RTL Design
IP Interfaces Hardware Architecture

Verification
Behavioral Simulation Functional Simulation Static Timing Analysis Timing Simulation

Implement Design FPGA Requirements

Hardware Specification Place & Route Test Stimulus Synthesis Map

Back Annotation

FPGA Hardware

Model-Based Design for FPGAs

MATLAB and Simulink Algorithm and System Design
RTL Design
IP Interfaces

Verification
Behavioral Simulation Functional Simulation Static Timing Analysis Timing Simulation

Automatic HDL Code Generation

Hardware Architecture

RTL

Implement Design
Synthesis Map Place & Route

Back Annotation

FPGA Hardware

Model-Based Design for FPGAs

MATLAB and Simulink Algorithm and System Design
Verification
Behavioral Simulation

Automatic HDL Code Generation

Behavioral Cosimulation

Functional Simulation Static Timing Analysis Timing Simulation

RTL

Implement Design
Synthesis Map Place & Route

Back Annotation

FPGA Hardware

Model-Based Design for FPGAs

MATLAB and Simulink Algorithm and System Design
Verification

Automatic HDL Code Generation

Behavioral Cosimulation

Functional Simulation Static Timing Analysis Timing Simulation

RTL

Implement Design
Back Annotation Synthesis Map

Back Annotation

Implement Design
Synthesis
Map Place & Route

Verification
Functional Simulation Static Timing Analysis Timing Simulation

Place & Route

FPGA Hardware

Model-Based Design for FPGAs

MATLAB and Simulink Algorithm and System Design

Automatic HDL Code Generation

Behavioral Cosimulation

RTL

Back Annotation

Implement Design
Synthesis
Map Place & Route

Verification
Functional Simulation Static Timing Analysis Timing Simulation

FPGA Hardware

FPGA Hardware
7

Model-Based Design with FPGA Workflows

MathWorks
HDL Coder

Altera
DSP Builder DSP Builder Advanced Blockset

HDL Coder

HDL code generation from MATLAB code, Simulink models, and Stateflow charts
Target-independent HDL code IEEE 1376 compliant VHDL IEEE 1364-2001 compliant Verilog

MATLAB, Simulink , Stateflow Algorithm and System Design

HDL Coder Link for ModelSim HDL Verifier

Verification
Generate HDL test bench Cosimulate with Mentor Graphics ModelSim or Cadence Incisive

Workflow integration
Integration with FPGA implementation tools

e.g., Quartus II

Optimize for area/speed Generate bit stream Back-annotation of timing to Simulink model

HDL

FPGA

Verify
ASIC
9

Generate

Digital Down Converter

MATLAB/Simulink Algorithm Model

DDC accepts
A high sample rate passband signal (may be 50 to 100 Msps)

DDC produces
A low sample-rate baseband signal ready for demodulation
Fixed-Point Model

HW Implementation Model

HDL

10

Fixed-Point Analysis
Digital Down Converter

Convert floating-point to fixed-point models

Automatic tracking of signal range (also intermediate quantities) Fraction lengths recommendation

Bit-true models in the same environment

Quantify the impact of fixed point quantization

Find and fix issues with fixed point easily

11

Automatic HDL Code Generation

Digital Down Converter
Automatically generate bit true, cycle accurate HDL code from Simulink, MATLAB, and Stateflow

Full traceability between model and HDL code

12

Simulink Library Support for HDL

HDL Supported Blocks

Over 170 blocks supported Core Simulink Blocks

Basic and Array Arithmetic, Look-Up Table, Signal Routing (Mux/Demux, Delay, Selector), Logic & Bit Operations, Dual and single port RAM, FIFO, CORDIC

Signal Processing Blocks

NCO, FFT, Digital Filters (FIR, IIR, Multi-rate, Adaptive), Rate Change (Up-/Down-sample), Statistics (Min/Max)

Communications Blocks
Psuedo-random Sequence Generator, Modulator / Demodulator, Interleaver / Deinterleaver, Viterbi Decoder

13

MATLAB and Stateflow

HDL Supported Blocks

MATLAB Function Block

Subset of the MATLAB language for modeling and generating HDL implementations MATLAB Function Block Design Patterns for HDL

Stateflow
Tool for graphical modeling of Mealy / Moore finite state machines

14

Use Cases for HDL Coder

System-level designers new to hardware design

Companies/design teams that have adopted Model-Based Design

MATLAB, Simulink , Stateflow Algorithm and System Design

HDL Coder Link for ModelSim HDL Verifier

Target-independent HDL is needed

Safety-critical applications (DO-254) Retargeting algorithm between software and hardware Model includes MATLAB code or Stateflow charts

HDL

FPGA

Verify
ASIC

Generate

16

Altera DSP Builder

Generates HDL optimized for Altera FPGAs from Simulink

Key Optimizations
Silicon Architecture Aware Constraint Driven Design

Simulink Algorithm and System Design

Link forBuilder DSP ModelSim

Generate

Single Precision Double Precision

Verification
Generate HDL test bench Cosimulate with ModelSim
Altera FPGA FPGA

HDL HDL Altera ASIC HardCopy ASIC

Verify

Supports Floating Point

17

Altera DSP Builder

Simulink
Sources Sinks Continuous Discrete Nonlinear Math

Altera DSP Builder Advanced Blockset Altera DSP Builder Blockset Fixed-Point Blockset DSP System Blockset Simulink Coder Communications System Blockset Image Acquisition Toolbox others
18

Altera DSP Builder Design Flow

MATLAB / Simulink (System simulation and verification)

HDL / Hardware Domain ( Hardware implementation / RTL simulation)

19

Constraint-Driven Design (1)

1. Choose your type Fixed or Floating Point

20

Constraint-Driven Design (2)

2. Create Model

Use ModelIP or ModelPrim libraries

ModelIP

ModelPrim

21

Constraint-Driven Design (3)

3. Select Device

Device-independent modeling up to this level

Stratix II AUTO

22

Constraint-Driven Design (4)

4. Set Frequency

Automatic pipelining or time sharing (ModelIP)

23

Constraint-Driven Design (5)

5. Compile

24

Math.h Functions

Math.h
- SIN - COS - TAN - ASIN - ACOS - ATAN - EXP - LOG - LOG10 - POW(x,y) - LDEXP - FLOOR - CEIL - FABS - SQRT - DIVIDE - 1/SQRT

Implemented in Floating Point

25

Complex Functions Easily Realized in FPGA

26

Supports Single and Double Precision

Including complex (c)

27

Supports Fixed Point and Floating Point

Within the same model

28

Loop Block for For i Loops

29

Nested Looping Functionality

30

Use Cases Summary

HDL Coder

Altera DSP Builder

System-level designers new to hardware design Companies that have adopted Model-Based Design Target-independent HDL needed Safety-critical applications

Hardware designers new to Model-Based Design Optimized for all Altera devices even ones we havent built yet Constraint driven for high performance designs

Retargeting algorithm between software and hardware

Model includes MATLAB code or Stateflow charts

Supports floating-point implementation

Supports integration into larger systems with Altera Qsys

31

Next Steps
1.

Visit mathworks.com/fpga
Get free FPGA Design Information Kit

View webinar Accelerate FPGA Design Using Simulink HDL Coder

Request trial of HDL Coder

2.

Visit altera.com/dspbuilder
Technical resources Request trial

3.

Questions?
Write to [email protected]

MATLAB and Simulink are registered trademarks of The MathWorks, Inc. See www.mathworks.com/trademarks for a list of additional trademarks. Other product or brand names may be trademarks or registered trademarks of their respective holders 32