CULA in the kernel function of CUDA

[thulzx]

Hi,
I am using CULA function to accelerate my algorithm.

Can the CULA function be used in the kernel function of CUDA?

For example, I use the function 'culaDeviceDposv' in the following kernel function

__global__ void vvdot(double *a1, double *a2, double *res)
{
...
culaDeviceDposv('U',10,10,a1,10,a2,10);
...
}

The VS compiler tells me that the culaDeviceDposv can not be used in that way.

Does it mean that the CULA function can only be used in the host?

Thank you very much.

Best regards,
thulzx

[john]

CULA functions are host only.

[thulzx]

[thulzx]