Hi all,
I implemented fast matrix multiplication in GroovyLab using OpenBLAS as:
final public Matrix multiply( double[][] that) {
double [] flmThis = Mat.oneDTransposeDoubleArray(d); // construct
a FloatMatrix for the receiver
double [] flmThat = Mat.oneDTransposeDoubleArray(that); //
construct a FloatMatrix for the argument
int Arows = Nrows(); int Acolumns = Ncols();
int Ccolumns = that[0].length;
double [] result = new double[Arows*Ccolumns];
double alpha=1.0;
double beta=0.0;
int lda = Arows;
int ldb = Acolumns;
int ldc = Arows;
// perform the multiplication using openblas
cblas_dgemm(CblasColMajor, CblasNoTrans, CblasNoTrans, Arows,
Ccolumns, Acolumns, alpha, flmThis, lda, flmThat, ldb, beta,
result, ldc);
double [][] rd = new double[Arows][Ccolumns];
int cnt = 0;
int r = 0;
int c = 0;
while (c < Ccolumns) {
r=0;
while (r < Arows ) {
rd[r][c] = result[cnt];
cnt++;
r++;
}
c++;
}
return new Matrix(rd);
}
public Matrix multiply( Matrix that) {
return multiply(that.d);
}
however, calling it on Matrix objects is slow:
n=2000
x=rand(n,n)
tic()
xx=x*x // Matrix * Matrix, slow however the method that calls OpenBLAS
is called
tm=toc() // delays a lot!!
but however on double[][] * Matrix is very fast:
n=2000
x=rand(n,n)
tic()
xx=x.d * x // double[][] * Matrix, is very fast !!
tm=toc() // very fast, uses OpenBLAS
What can happen here?
Best Regards
Stergios
