topic Thanks a lot! in Intel® oneAPI Math Kernel Library

Use array of floats as argument

Carlos_Alexandre_T_ — Wed, 26 Feb 2014 00:28:08 GMT

Is it possible to use an array of floats as parameters when using dgemm?

Ying_H_Intel — Wed, 26 Feb 2014 01:03:35 GMT

Hello Carlos,

MKL function allows float array, either double or single floats as input. for example,

double A[m*k];

B = (double*) malloc(sizeof(double)*k*n);
C = (double*) malloc(sizeof(double)*m*n);

dgemm(&transa, &transb, &m, &n, &k, &alpha, A, &m, B, &k, &beta, C, &m);

Moreover, the first letter in *gemm indicate the data type. and you can see many sample codes under <MKL install dir>/examples/

for example, C:\Program Files (x86)\Intel\Composer XE 2013\mkl\examples\cblas\source

s real, single precision
c complex, single precision
d real, double precision
z complex, double precision

Best Regards,

Ying

P.S it seems the post was showned many times, we will delete the rest of them.

Carlos_Alexandre_T_ — Wed, 26 Feb 2014 02:02:34 GMT

Thanks a lot!

Did you see this under MKL's Documentation?

Ying_H_Intel — Wed, 26 Feb 2014 07:08:21 GMT

Hi Carlos,

Right, MKL Documetation(MKL manual and MKL userguide) includes the informations.

The doc can be found in MKL install directory or from on-line documentation: http://software.intel.com/en-us/intel-software-technical-documentation

Best Regards,

Ying