dpctl.SyclDevice.preferred_vector_width_double

SyclDevice.preferred_vector_width_double

Returns the preferred native vector width size for built-in scalar types that can be put into vectors.

Returns:

Preferred native vector width size for C type double.

Return type:

int

If device does not support double-precision floating point operations, the native width is zero.

Example:

import dpctl

dev = dpctl.select_cpu_device()
pvw_d = dev.preferred_vector_width_double