Listing 2. Simplified Matrix Multiplication in Brook

kernel void
matmulKernel( float N2, float A[][], float B[][],
                          out float result<> ){
  float2 ik = indexof(result).xy;
  float4 ijjk = float4( ik.x, 0.0f, 0.0f, ik.y );
  float4 jp1 = float4( 0.0f, 1.0f, 1.0f, 0.0f );

  float C = 0.0f;
  float n2 = N2;

  while( n2 > 0 ) {
    C += A[ijjk.zw]*B[ijjk.xy];
    ijjk += jp1;
    n2 -= 1.0f;
  }
  result = C;
}


void
matmul( float* A, float* B, float* C,
            size_t N1, size_t N2, size_t N3 ){
  float Astream<N1, N2>;
  float Bstream<N2, N3>;
  float Cstream<N1, N3>;

  streamRead( Astream, A );
  streamRead( Bstream, B );
  matmulKernel( (float)N2, Astream, Bstream, Cstream );

  streamWrite( Cstream, C );
}