wiesendsi102436/info2Praktikum-NeuronalesNetz
1
0
Fork 0
forked from freudenreichan/info2Praktikum-NeuronalesNetz
Code Pull Requests Activity

make matrix multiplication faster

Browse Source
This commit is contained in:
Simon Wiesend 2025-11-23 16:31:18 +01:00
parent 00ac1aa04a
commit adf75b66d2
Signed by: wiesendsi102436
GPG Key ID: C18A833054142CF0
1 changed files with 17 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
matrix.c
View File

@ -109,7 +109,6 @@ Matrix add(const Matrix matrix1, const Matrix matrix2)
{
for (size_t n = 0; n < matrix1.cols; n++)
{
// this is unnecessarily complicated because at this point we already know that the matrices are compatible
setMatrixAt(getMatrixAt(matrix1, m, n) + getMatrixAt(matrix2, m, n), resMat, m, n);
}
}
@ -117,32 +116,37 @@ Matrix add(const Matrix matrix1, const Matrix matrix2)
return resMat;
}
Matrix multiply(const Matrix matrix1, const Matrix matrix2)
Matrix multiply(const Matrix A, const Matrix B)
{
if (matrix1.cols != matrix2.rows || matrix1.buffer == NULL || matrix2.buffer == NULL)
if (A.cols != B.rows || A.buffer == NULL || B.buffer == NULL)
{
return createMatrix(0, 0);
}
int rows = matrix1.rows, cols = matrix2.cols;
Matrix resMat = createMatrix(rows, cols);
int rows = A.rows, cols = B.cols;
Matrix C = createMatrix(rows, cols);
if (resMat.buffer == NULL)
if (C.buffer == NULL)
{
return createMatrix(0, 0);
}
for (size_t rowIdx = 0; rowIdx < rows; rowIdx++)
// M = Rows, K = Common Dim, N = Cols
size_t M = A.rows, K = A.cols, N = B.cols;
for (size_t i = 0; i < M; i++)
{
for (size_t colIdx = 0; colIdx < cols; colIdx++)
for (size_t k = 0; k < K; k++)
{
int curCellVal = 0;
for (size_t k = 0; k < matrix1.cols; k++)
MatrixType valA = A.buffer[i * K + k];
for (size_t j = 0; j < N; j++)
{
curCellVal += getMatrixAt(matrix1, rowIdx, k) * getMatrixAt(matrix2, k, colIdx);
}
setMatrixAt(curCellVal, resMat, rowIdx, colIdx);
// C[i, j] += A[i, k] * B[k, j];
// M x N, M x K, K x N
C.buffer[i * N + j] += valA * B.buffer[k * N + j];
}
}
return resMat;
}
return C;
}