0

I'm attempting to use cblas_dgemm to perform matrix multiplication on a 1 x N and an N x N matrix. My call is as follows:

cblas_dgemm(CblasRowMajor, CblasNoTrans, CblasNoTrans,
    1, width, width, 1.0, A, width, B, width,
    0.0, C, width);

however what I get in C is simply a copy of A.

Here is the documentation for the method: https://software.intel.com/en-us/node/429920

I've gone through the parameters a number of times and they all seem fine so I was wondering if anyone could pinpoint something stupid I'm doing?

James Bilous
  • 167
  • 9
  • 2
    I don't see anything wrong with your parameters. Perhaps you could provide this code snippet in context, with definitions of A & B? – newling Nov 15 '15 at 08:23

1 Answers1

0

I simply needed to transpose matrix B or set CblasColMajor since my matrix B was already stored in col major order.

James Bilous
  • 167
  • 9