Implement Hardware-Efficient Real Partial-Systolic Matrix Solve Using Q-less QR Decomposition

Open Script

This example shows how to implement a hardware-efficient solution to the real-valued matrix equation A'AX=B using the Real Partial-Systolic Matrix Solve Using Q-less QR Decomposition block.

Define Matrix Dimensions

Specify the number of rows in matrix A, the number of columns in matrix A and rows in B, and the number of columns in matrix B.

m = 300; % Number of rows in A
n = 10;  % Number of columns in A and rows in B
p = 1;   % Number of columns in B

Generate Matrices

For this example, use the helper function realRandomQlessQRMatrices to generate random matrices A and B for the problem A'AX=B. The matrices are generated such that the elements of A and B are between -1 and +1, and A is full rank.

rng('default')
[A,B] = fixed.example.realRandomQlessQRMatrices(m,n,p);

Select Fixed-Point Data Types

Use the helper function realQlessQRMatrixSolveFixedpointTypes to select fixed-point data types for input matrices A and B, and output X such that there is a low probability of overflow during the computation.

max_abs_A = 1;      % Upper bound on max(abs(A(:))
max_abs_B = 1;      % Upper bound on max(abs(B(:))
precisionBits = 24; % Number of bits of precision
T = fixed.realQlessQRMatrixSolveFixedpointTypes(m,n,max_abs_A,max_abs_B,precisionBits);
A = cast(A,'like',T.A);
B = cast(B,'like',T.B);
OutputType = fixed.extractNumericType(T.X);

Open the Model

model = 'RealPartialSystolicQlessQRMatrixSolveModel';
open_system(model);

The Data Handler subsystem in this model takes real matrices A and B as inputs. The readyA and readyB ports trigger the Data Handler. After sending a true validIn signal, there may be some delay before ready is set to false. When the Data Handler detects the leading edge of the readyA signal, the block sets validInA to true and sends the next row of A. When the Data Handler detects the leading edge of the readyB signal, the block sets validInB to true and sends the next matrix B. This protocol allows data to be sent whenever a leading edge of a ready signal is detected, ensuring that all data is processed.

Set Variables in the Model Workspace

Use the helper function setModelWorkspace to add the variables defined above to the model workspace. These variables correspond to the block parameters for the Real Partial-Systolic Matrix Solve Using Q-less QR Decomposition block.

numSamples = 1; % Number of samples
fixed.example.setModelWorkspace(model,'A',A,'B',B,'m',m,'n',n,'p',p,...
    'regularizationParameter',0,...
    'numSamples',numSamples,'OutputType',OutputType);

Simulate the Model

out = sim(model);

Construct the Solution from the Output Data

The Real Partial-Systolic Matrix Solve Using Q-less QR Decomposition block outputs matrix X at each time step. When a valid result matrix is output, the block sets validOut to true.

X = out.X;

Verify the Accuracy of the Output

To evaluate the accuracy of the Real Partial-Systolic Matrix Solve Using Q-less QR Decomposition block, compute the relative error.

relative_error = norm(double(A'*A*X - B))/norm(double(B)) %#ok<NOPTS>

relative_error =

   8.6289e-05