MathWorks Fixed Point Team

MathWorks

Active since 2017

Followers: 0 Following: 0

Message

The engineering team develop Fixed-Point Designer product at MathWorks. Share best practices to embedded numerical efficiency in Model-Based Design in MATLAB and Simulink.

Statistics

All

View badges

Feeds

Answered
Why am I getting "double" datatype automatically for every simulink block?
Data Type Override setting on the model or selected subsystem is the likely culpit for this behavior. Use get_param API's to q...

ongeveer 2 maanden ago | 1

Answered
What are the efficient ways to generate “mod by constant” code with Embedded Coder (e.g., mod(u, 360))?
If your goal is to generate efficient, inlined code for mod by a known constant, use the Modulo by Constant block rather than th...

ongeveer 2 maanden ago | 0

Question

What are the efficient ways to generate “mod by constant” code with Embedded Coder (e.g., mod(u, 360))?
I’m looking for an efficient way to compute mod by a compile-time constant in Simulink models that generate C/C++ with Embedded ...

ongeveer 2 maanden ago | 1 answer | 0

1

answer

Answered
How does the Compare To Constant block in Simulink choose type of the Constant value, and what are the implications for fixed-point operations?
The Compare To Constant block in Simulink casts the Constant value parameter to the input signal's data type before performing t...

2 maanden ago | 1

| accepted

Question

How does the Compare To Constant block in Simulink choose type of the Constant value, and what are the implications for fixed-point operations?
How can I detect and address questionable fixed-point operations in Compare To Constant block? This example model generates co...

2 maanden ago | 1 answer | 0

1

answer

Answered
Avoid autoscaling of wordlength and fractionlength in fixed point objects
SQRT function computes the square root of a fi object using a bisection algorithm. You can use additional numerictype to contro...

3 maanden ago | 0

Answered
How can I use bit operation functions like bitconcat with both fi and built-in integer types without branching on isfi?
Use castIntToFi to handle both fi and integer types. If you want your code to work with both built-in integer types and fi obje...

3 maanden ago | 0

| accepted

Question

How can I use bit operation functions like bitconcat with both fi and built-in integer types without branching on isfi?
I'm working with bit operation functions (e.g., bitconcat) that only accept fi objects as input. My code sometimes deals with bu...

3 maanden ago | 1 answer | 0

1

answer

Answered
How do I debug "Saturate on overflow detected" warnings from Simulink blocks?
This warning (can be reported as error) indicates that the output signal from this block has exceeded the representable range of...

3 maanden ago | 0

| accepted

Question

How do I debug "Saturate on overflow detected" warnings from Simulink blocks?
I noticed the warning messages like this after running the simulation: Saturate on overflow detected. This originated from 'mod...

3 maanden ago | 1 answer | 0

1

answer

Answered
Why do the values of learnables in a quantized dlnetwork still stored as float32(single precision)?
Yes, the learnables on the dlnetwork/quantized network are still stored as single precision. Consider estimating parameter mem...

4 maanden ago | 0

| accepted

Question

Why do the values of learnables in a quantized dlnetwork still stored as float32(single precision)?
Even though the dlquantizer is quantizing the weights of the fully connected layer to int8 and bias of the layer to int32, why d...

4 maanden ago | 1 answer | 0

1

answer

Answered
Why did my single precision model generate double data type sine functions, using sin(real_T) instead of sinf(real32_T)?
You can try to run Single Precision Converter, https://www.mathworks.com/help/fixedpoint/ug/getting-started-with-the-double-to-s...

4 maanden ago | 1

| accepted

Question

Why did my single precision model generate double data type sine functions, using sin(real_T) instead of sinf(real32_T)?
I am trying to generate code for TI target. My model has sine block for single data type signals. I found generated code have co...

4 maanden ago | 1 answer | 0

1

answer

Submitted

Deep Learning Toolbox Model Compression Library
Optimize deep learning models with efficient compression techniques

11 maanden ago | 53 downloads |

3.8 / 5

Question

FAQ: Why does Fixed-Point fi variable times MATLAB floating-point variable error out in code generation, MATLAB Function Block, etc.?
FAQ: Why does Fixed-Point fi variable multiplied by a MATLAB floating-point variable error out in code generation, MATLAB Functi...

bijna 2 jaar ago | 1 answer | 0

1

answer

Question

FAQ: What's the difference between the two Simulink Precision Loss Diagnostics for Parameters and Fixed-Point Constants?
What is the difference between the Detect precision loss that applies to fixed-point constants and the Detect precision loss for...

bijna 2 jaar ago | 1 answer | 0

1

answer

Question

FAQ: How can I resolve the error due to lookup table not allowing repeated breakpoints?
I'm trying to model abrupt discontinuities using lookup tables. For example, consider this data that has abrupt discontinuities...

meer dan 2 jaar ago | 1 answer | 0

1

answer

Question

FAQ: Why does a value look more precise when displayed in MATLAB or Simulink than in the generated code?
A value used in my design looks more precise when displayed in MATLAB or Simulink. The same value in C or C++ generated by Embe...

bijna 3 jaar ago | 1 answer | 0

1

answer

Answered
FAQ: Why is a fixed-point type's Fraction Length or Integer Length sometimes negative?
Maximizing precision for limited bits The ability to have a negative fraction lengths or a negative integer lengths is a very g...

bijna 3 jaar ago | 0

| accepted

Question

FAQ: Why is a fixed-point type's Fraction Length or Integer Length sometimes negative?
Sometimes fixed-point variables have fraction lengths that are negative. a = 5.6632765314184e+15 aFi = fi(a,1,4) curFractionL...

bijna 3 jaar ago | 1 answer | 0

1

answer

Question

Why does fi signed plus unsigned produce needless range bits?
When adding signed and unsigned fi variables with the default full-precision settings, sometimes the output type has needless ra...

ongeveer 3 jaar ago | 1 answer | 0

1

answer

Question

How can I use int64 and uint64 with my Simulink S-function?
How can I use int64 and uint64 with my Simulink S-function?

meer dan 3 jaar ago | 1 answer | 0

1

answer

Question

Why does Simulink's Weighted Sample Time block sometimes use fixed-point data types for its output?
In Simulink, I want to get the sample time of a signal and use that value in other calculations. The Weighted Sample Time block ...

bijna 4 jaar ago | 1 answer | 0

1

answer

Question

How can I use fractional data types when the values are beyond the interval [-1, +1) ?
Many references describe implementing fixed-point designs using fractional data types. These signed fractional data types repres...

meer dan 4 jaar ago | 1 answer | 0

1

answer

Question

FAQ: Why aren't boolean and 1 bit unsigned integers treated as the same data type?
Variables and signals of the boolean data type (called logical in MATLAB) only represent the values 0 and 1. Variables and sign...

bijna 5 jaar ago | 1 answer | 0

1

answer

Question

Best practices for converting MATLAB code to fixed point
I have tried to use fixed-point data types (i.e. fi) for MATLAB code. Are there any best practices for this process?

ongeveer 5 jaar ago | 1 answer | 0

1

answer

Submitted

Fixed-Point Made Easy for FPGA Programming
Material used in the "Fixed-Point Made Easy for FPGA Programming" webinar.

ongeveer 5 jaar ago | 1 download |

4.0 / 5

Question

Some forms of Simulink logging write 64 bit integers as fi objects. Does that mean I need a Fixed-Point Designer license?
I've observed that some forms of logging in Simulink write integer variables out to MATLAB as fi objects. For example, DSP Bina...

ongeveer 5 jaar ago | 1 answer | 0

1

answer

Answered
How can I use custom constraints for fxpopt?
fxpopt workflow allows user to automatically assign fixed-point data types on a system by trying a basic trial and error loop. ...

ongeveer 5 jaar ago | 0

| accepted