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

MOVNTPD

Moves a packed double-precision floating-point value from a register to a memory location using non-temporal hints, which bypasses the cache hierarchy to minimize cache pollution.

The table below covers the supported source and destination operands.

DO NOT support LOCK

This instruction SHALL only be used when the destination memory address is aligned on a 16-byte boundary. If the address is not aligned, an exception may occur or performance may be severely degraded depending on the processor implementation.

The use of non-temporal stores implies that the data is not expected to be reused soon. Consequently, this instruction DOES NOT maintain memory ordering with respect to other stores. A SFENCE instruction SHALL be executed to ensure that all non-temporal stores are globally visible before subsequent memory operations that depend on the stored data. Failure to use a memory fence may result in data inconsistency in multi-processor systems.