Compare commits

9 Commits

Author SHA1 Message Date
48b016d8b7 Merge pull request 'Updating the readme' (#7) from update-readme into main
Reviewed-on: #7
2025-06-30 19:05:53 +00:00
8e4595f2ef Updated readme
All checks were successful
Merge-Checker / build_and_test (pull_request) Successful in 1m10s
2025-06-30 14:52:48 -04:00
99c0d3ed70 Merge pull request 'Adjusted timing test repetition and added QR decomposition' (#6) from Minor-cicd-fixes into main
All checks were successful
Merge-Checker / build_and_test (pull_request) Successful in 1m13s
Reviewed-on: #6
2025-06-10 23:06:02 +00:00
80c4ebfece Put time usage back
All checks were successful
Merge-Checker / build_and_test (pull_request) Successful in 1m18s
2025-06-07 11:08:56 -04:00
8b6f1de822 Updated timing test timings
All checks were successful
Merge-Checker / build_and_test (pull_request) Successful in 1m17s
2025-06-07 11:03:55 -04:00
719fc4d28a Adjusted timing test repetition and added QR decomposition
Some checks failed
Merge-Checker / build_and_test (pull_request) Has been cancelled
2025-06-07 10:58:59 -04:00
2a7eb93ebe Merge pull request 'Working on adding efficient eigenvector and value calculations' (#2) from eigenvector-and-values into main
Reviewed-on: #2
2025-06-06 22:32:18 +00:00
c099dfe760 Throwing in the towel on eigenvectors for now
All checks were successful
Merge-Checker / build_and_test (pull_request) Successful in 26s
2025-06-06 16:33:20 -04:00
1715d2b46c Merge pull request 'Add a merge checker script' (#1) from Testing-merge-checker into main
Reviewed-on: #1
2025-05-29 20:36:30 +00:00
5 changed files with 97 additions and 128 deletions

View File

@@ -2,8 +2,11 @@
This matrix math library is focused on embedded development and avoids any heap memory allocation unless you explicitly ask for it.
It uses templates to pre-allocate matrices on the stack.
There are still several operations that are works in progress such as:
- Add a function to calculate eigenvalues/vectors
- Add a function to compute RREF
- Add a function for SVD decomposition
- Add a function for LQ decomposition
# Building
1. Initialize the repositiory with the command:
```bash
cmake -S . -B build -G Ninja
```
2. Go into the build folder and run `ninja`
3. That's it. You can test out the build by running `./unit-tests/matrix-tests`

View File

@@ -572,12 +572,13 @@ void Matrix<rows, columns>::EigenQR(Matrix<rows, rows> &eigenVectors,
Matrix<rows, rows> Ak = *this; // Copy original matrix
Matrix<rows, rows> QQ{Matrix<rows, rows>::Identity()};
Matrix<rows, rows> shift{0};
for (uint32_t iter = 0; iter < maxIterations; ++iter) {
Matrix<rows, rows> Q, R, shift;
Matrix<rows, rows> Q, R;
// QR shift lets us "attack" the first diagonal to speed up the algorithm
shift = Matrix<rows, rows>::Identity() * Ak[rows - 1][rows - 1];
// // QR shift lets us "attack" the first diagonal to speed up the algorithm
// shift = Matrix<rows, rows>::Identity() * Ak[rows - 1][rows - 1];
(Ak - shift).QRDecomposition(Q, R);
Ak = R * Q + shift;
QQ = QQ * Q;

View File

@@ -559,28 +559,6 @@ TEST_CASE("QR Decompositions", "Matrix") {
REQUIRE(std::fabs(R[i][j]) < 1e-4f);
}
}
// check that all Q values are correct
REQUIRE_THAT(Q[0][0], Catch::Matchers::WithinRel(0.1231f, 1e-4f));
REQUIRE_THAT(Q[0][1], Catch::Matchers::WithinRel(0.904534f, 1e-4f));
REQUIRE_THAT(Q[0][2], Catch::Matchers::WithinRel(0.0f, 1e-4f));
REQUIRE_THAT(Q[1][0], Catch::Matchers::WithinRel(0.49237f, 1e-4f));
REQUIRE_THAT(Q[1][1], Catch::Matchers::WithinRel(0.301511f, 1e-4f));
REQUIRE_THAT(Q[1][2], Catch::Matchers::WithinRel(0.0f, 1e-4f));
REQUIRE_THAT(Q[2][0], Catch::Matchers::WithinRel(0.86164f, 1e-4f));
REQUIRE_THAT(Q[2][1], Catch::Matchers::WithinRel(-0.30151f, 1e-4f));
REQUIRE_THAT(Q[2][2], Catch::Matchers::WithinRel(0.0f, 1e-4f));
// check that all R values are correct
REQUIRE_THAT(R[0][0], Catch::Matchers::WithinRel(8.124038f, 1e-4f));
REQUIRE_THAT(R[0][1], Catch::Matchers::WithinRel(9.60114f, 1e-4f));
REQUIRE_THAT(R[0][2], Catch::Matchers::WithinRel(11.07823f, 1e-4f));
REQUIRE_THAT(R[1][0], Catch::Matchers::WithinRel(0.0f, 1e-4f));
REQUIRE_THAT(R[1][1], Catch::Matchers::WithinRel(0.90453f, 1e-4f));
REQUIRE_THAT(R[1][2], Catch::Matchers::WithinRel(1.80907f, 1e-4f));
REQUIRE_THAT(R[2][0], Catch::Matchers::WithinRel(0.0f, 1e-4f));
REQUIRE_THAT(R[2][1], Catch::Matchers::WithinRel(0.0f, 1e-4f));
REQUIRE_THAT(R[2][2], Catch::Matchers::WithinRel(0.0f, 1e-4f));
}
SECTION("4x2 QRDecomposition") {
@@ -622,42 +600,41 @@ TEST_CASE("QR Decompositions", "Matrix") {
}
}
// TEST_CASE("Eigenvalues and Vectors", "Matrix") {
// SECTION("2x2 Eigen") {
// Matrix<2, 2> A{1.0f, 2.0f, 3.0f, 4.0f};
// Matrix<2, 2> vectors{};
// Matrix<2, 1> values{};
TEST_CASE("Eigenvalues and Vectors", "Matrix") {
SECTION("2x2 Eigen") {
Matrix<2, 2> A{1.0f, 2.0f, 3.0f, 4.0f};
Matrix<2, 2> vectors{};
Matrix<2, 1> values{};
// A.EigenQR(vectors, values, 1000000, 1e-20f);
A.EigenQR(vectors, values, 1000000, 1e-20f);
// REQUIRE_THAT(vectors[0][0], Catch::Matchers::WithinRel(0.41597f, 1e-4f));
// REQUIRE_THAT(vectors[1][0], Catch::Matchers::WithinRel(0.90938f, 1e-4f));
// REQUIRE_THAT(values[0][0], Catch::Matchers::WithinRel(5.372282f, 1e-4f));
// REQUIRE_THAT(values[1][0], Catch::Matchers::WithinRel(-0.372281f,
// 1e-4f));
// }
REQUIRE_THAT(vectors[0][0], Catch::Matchers::WithinRel(0.41597f, 1e-4f));
REQUIRE_THAT(vectors[1][0], Catch::Matchers::WithinRel(0.90938f, 1e-4f));
REQUIRE_THAT(values[0][0], Catch::Matchers::WithinRel(5.372282f, 1e-4f));
REQUIRE_THAT(values[1][0], Catch::Matchers::WithinRel(-0.372281f, 1e-4f));
}
// SECTION("3x3 Eigen") {
// // this symmetrix tridiagonal matrix is well behaved for testing
// Matrix<3, 3> A{1, 2, 3, 4, 5, 6, 7, 8, 9};
SECTION("3x3 Rank Defficient Eigen") {
SKIP("Skipping this because QR decomposition isn't ready for it");
// this symmetrix tridiagonal matrix is well behaved for testing
Matrix<3, 3> A{1, 2, 3, 4, 5, 6, 7, 8, 9};
// Matrix<3, 3> vectors{};
// Matrix<3, 1> values{};
// A.EigenQR(vectors, values, 1000000, 1e-8f);
Matrix<3, 3> vectors{};
Matrix<3, 1> values{};
A.EigenQR(vectors, values, 1000000, 1e-8f);
// std::string strBuf1 = "";
// vectors.ToString(strBuf1);
// std::cout << "Vectors:\n" << strBuf1 << std::endl;
// strBuf1 = "";
// values.ToString(strBuf1);
// std::cout << "Values:\n" << strBuf1 << std::endl;
std::string strBuf1 = "";
vectors.ToString(strBuf1);
std::cout << "Vectors:\n" << strBuf1 << std::endl;
strBuf1 = "";
values.ToString(strBuf1);
std::cout << "Values:\n" << strBuf1 << std::endl;
// REQUIRE_THAT(vectors[0][0], Catch::Matchers::WithinRel(0.23197f, 1e-4f));
// REQUIRE_THAT(vectors[1][0], Catch::Matchers::WithinRel(0.525322f,
// 1e-4f)); REQUIRE_THAT(vectors[2][0], Catch::Matchers::WithinRel(0.81867f,
// 1e-4f)); REQUIRE_THAT(values[0][0], Catch::Matchers::WithinRel(-1.11684f,
// 1e-4f)); REQUIRE_THAT(values[1][0], Catch::Matchers::WithinRel(0.0f,
// 1e-4f)); REQUIRE_THAT(values[2][0], Catch::Matchers::WithinRel(16.1168f,
// 1e-4f));
// }
// }
REQUIRE_THAT(vectors[0][0], Catch::Matchers::WithinRel(0.23197f, 1e-4f));
REQUIRE_THAT(vectors[1][0], Catch::Matchers::WithinRel(0.525322f, 1e-4f));
REQUIRE_THAT(vectors[2][0], Catch::Matchers::WithinRel(0.81867f, 1e-4f));
REQUIRE_THAT(values[0][0], Catch::Matchers::WithinRel(-1.11684f, 1e-4f));
REQUIRE_THAT(values[1][0], Catch::Matchers::WithinRel(0.0f, 1e-4f));
REQUIRE_THAT(values[2][0], Catch::Matchers::WithinRel(16.1168f, 1e-4f));
}
}

View File

@@ -8,6 +8,7 @@
// any other libraries
#include <array>
#include <cmath>
#include <cstdint>
// basically re-run all of the matrix tests with huge matrices and time the
// results.
@@ -29,13 +30,13 @@ TEST_CASE("Timing Tests", "Matrix") {
Matrix<4, 4> mat5{};
SECTION("Addition") {
for (uint32_t i{0}; i < 10000; i++) {
for (uint32_t i{0}; i < 100000; i++) {
mat3 = mat1 + mat2;
}
}
SECTION("Subtraction") {
for (uint32_t i{0}; i < 10000; i++) {
for (uint32_t i{0}; i < 100000; i++) {
mat3 = mat1 - mat2;
}
}
@@ -47,19 +48,19 @@ TEST_CASE("Timing Tests", "Matrix") {
}
SECTION("Scalar Multiplication") {
for (uint32_t i{0}; i < 10000; i++) {
for (uint32_t i{0}; i < 100000; i++) {
mat3 = mat1 * 3;
}
}
SECTION("Element Multiply") {
for (uint32_t i{0}; i < 10000; i++) {
for (uint32_t i{0}; i < 100000; i++) {
mat1.ElementMultiply(mat2, mat3);
}
}
SECTION("Element Divide") {
for (uint32_t i{0}; i < 10000; i++) {
for (uint32_t i{0}; i < 100000; i++) {
mat1.ElementDivide(mat2, mat3);
}
}
@@ -68,52 +69,59 @@ TEST_CASE("Timing Tests", "Matrix") {
// what about matrices of 0,0 or 1,1?
// minor matrix for 2x2 matrix
Matrix<49, 49> minorMat1{};
for (uint32_t i{0}; i < 10000; i++) {
for (uint32_t i{0}; i < 100000; i++) {
mat1.MinorMatrix(minorMat1, 0, 0);
}
}
SECTION("Determinant") {
for (uint32_t i{0}; i < 100000; i++) {
for (uint32_t i{0}; i < 1000000; i++) {
float det1 = mat4.Det();
}
}
SECTION("Matrix of Minors") {
for (uint32_t i{0}; i < 100000; i++) {
for (uint32_t i{0}; i < 1000000; i++) {
mat4.MatrixOfMinors(mat5);
}
}
SECTION("Invert") {
for (uint32_t i{0}; i < 100000; i++) {
for (uint32_t i{0}; i < 1000000; i++) {
mat5 = mat4.Invert();
}
};
SECTION("Transpose") {
for (uint32_t i{0}; i < 10000; i++) {
for (uint32_t i{0}; i < 100000; i++) {
mat3 = mat1.Transpose();
}
}
SECTION("Normalize") {
for (uint32_t i{0}; i < 10000; i++) {
for (uint32_t i{0}; i < 100000; i++) {
mat3 = mat1 / mat1.EuclideanNorm();
}
}
SECTION("GET ROW") {
Matrix<1, 50> mat1Rows{};
for (uint32_t i{0}; i < 1000000; i++) {
for (uint32_t i{0}; i < 100000000; i++) {
mat1.GetRow(0, mat1Rows);
}
}
SECTION("GET COLUMN") {
Matrix<50, 1> mat1Columns{};
for (uint32_t i{0}; i < 1000000; i++) {
for (uint32_t i{0}; i < 100000000; i++) {
mat1.GetColumn(0, mat1Columns);
}
}
SECTION("QR Decomposition") {
Matrix<50, 50> Q, R{};
for (uint32_t i{0}; i < 500; i++) {
mat1.QRDecomposition(Q, R);
}
}
}

View File

@@ -1,56 +1,36 @@
Randomness seeded to: 2444679151
0.180 s: Addition
0.180 s: Timing Tests
0.177 s: Subtraction
0.177 s: Timing Tests
1.868 s: Multiplication
1.868 s: Timing Tests
0.127 s: Scalar Multiplication
0.127 s: Timing Tests
0.173 s: Element Multiply
0.173 s: Timing Tests
0.178 s: Element Divide
0.178 s: Timing Tests
0.172 s: Minor Matrix
0.172 s: Timing Tests
0.103 s: Determinant
0.103 s: Timing Tests
0.411 s: Matrix of Minors
0.411 s: Timing Tests
0.109 s: Invert
0.109 s: Timing Tests
0.122 s: Transpose
0.122 s: Timing Tests
0.190 s: Normalize
0.190 s: Timing Tests
0.006 s: GET ROW
0.006 s: Timing Tests
0.235 s: GET COLUMN
0.235 s: Timing Tests
Running matrix-timing-tests with timing
Randomness seeded to: 3567651885
1.857 s: Addition
1.857 s: Timing Tests
1.788 s: Subtraction
1.788 s: Timing Tests
1.929 s: Multiplication
1.929 s: Timing Tests
1.268 s: Scalar Multiplication
1.268 s: Timing Tests
1.798 s: Element Multiply
1.798 s: Timing Tests
1.802 s: Element Divide
1.803 s: Timing Tests
1.553 s: Minor Matrix
1.554 s: Timing Tests
1.009 s: Determinant
1.009 s: Timing Tests
4.076 s: Matrix of Minors
4.076 s: Timing Tests
1.066 s: Invert
1.066 s: Timing Tests
1.246 s: Transpose
1.246 s: Timing Tests
2.284 s: Normalize
2.284 s: Timing Tests
0.606 s: GET ROW
0.606 s: Timing Tests
24.629 s: GET COLUMN
24.630 s: Timing Tests
3.064 s: QR Decomposition
3.064 s: Timing Tests
===============================================================================
test cases: 1 | 1 passed
assertions: - none -
Command being timed: "build/unit-tests/matrix-timing-tests -d yes"
User time (seconds): 4.05
System time (seconds): 0.00
Percent of CPU this job got: 100%
Elapsed (wall clock) time (h:mm:ss or m:ss): 0:04.05
Average shared text size (kbytes): 0
Average unshared data size (kbytes): 0
Average stack size (kbytes): 0
Average total size (kbytes): 0
Maximum resident set size (kbytes): 3200
Average resident set size (kbytes): 0
Major (requiring I/O) page faults: 184
Minor (reclaiming a frame) page faults: 171
Voluntary context switches: 1
Involuntary context switches: 26
Swaps: 0
File system inputs: 12
File system outputs: 1
Socket messages sent: 0
Socket messages received: 0
Signals delivered: 0
Page size (bytes): 4096
Exit status: 0