This content has not been validated; it was AI-generated following the x86-64 ISA Documentation generation process to maintain high quality. However, as no human has verified this data, errors may exist. If the accuracy of this data is critical, please consult the Intel SDM. The model used to generate this is gemma4:31b

VPERMT2PD

Permutes two double-precision floating-point values from two sources based on an immediate. It selects two double-precision values from two 128-bit source operands and writes them to the destination.

The table after the description covers what the source and destinations can be.

DO NOT support LOCK

This instruction is only available in 64-bit mode. It requires AVX support.

The immediate serves as an index to select the specific double-precision values: bits 0-1 select the first element and bits 2-3 select the second element. An index of 0 or 1 refers to the first source, while 2 or 3 refers to the second source. Using an immediate value outside the range of 0-15 is not possible as the immediate is encoded as a 4-bit value; however, the architectural mapping only utilizes 4 bits. Failure to align the m128 memory operand to a 16-byte boundary may result in performance degradation or a general protection fault depending on the alignment check setting.