vec_pext

Purpose

Copies some bits from each element of one vector as specified by a mask in another vector into the rightmost bits of the target vector.

Note: This built-in function is valid only when the -mcpu option is set to target Power10 processors.

Syntax

d=vec_pext(a,b)

Result and argument types

The following table describes the types of the returned value and the function arguments.

Table 1. Types of the returned value and function arguments
d	a	b
vector unsigned long long	vector unsigned long long	vector unsigned long long

Result value

The bit of each element of a, for which the corresponding bit of that element of b is the rightmost 1-bit, is copied into the rightmost bit position of the corresponding element of d. The bit of each element of a, for which the corresponding bit of that element of b is the next-rightmost 1-bit, is copied into the next-rightmost bit position of the corresponding element of d, and so forth until the element of b contains no more 1-bits. All bits in each element of d that are not copied from the corresponding element of a are set to zero.

Here is an example for doubleword inputs:

Argument	Element 0	Element 1
a	0x 0088 0080 0008 0000	0x 4c48 4440 0c08 0400
b	0x 8888 8888 8888 8888	0x cccc cccc cccc cccc
d	0x 0000 0000 0000 3210	0x 0000 0000 7654 3210

Consider element 0 in the previous table. Since b contains 16 bits that are set to 1, all bits except the rightmost 16 bits of d are set to zero. The rightmost 16 bits are set by extracting the bits of a at the positions of the 1-bits in b. Those bits are 0011 0010 0001 0000, providing the result shown in the previous table.