Expect to add more x64 ISA about AI

[whyb]

AVX512(Intel® Advanced Vector Extensions 512)

AVX512_BITALG (expect to add) AVX512_VPOPCNTDQ (expect to add) AVX512_VP2INTERSECT (expect to add)

AVX512PF (Intel® Xeon Phi™ only.) AVX512ER (Intel® Xeon Phi™ only.) AVX512_4VNNIW (Intel® Xeon Phi™ only.) AVX512_4FMAPS (Intel® Xeon Phi™ only.)

AVX (VEX-encoded) versions of the Vector Neural Network Instructions

AVX512_VNNI (already exists) AVX_VNNI (already exists) AVX_VNNI_INT8 (already exists) AVX_VNNI_INT16 (expect to add)

Intel® AMX(Intel® Advanced Matrix Extensions)

AMX_FP16 (expect to add) AMX_COMPLEX (expect to add) AMX_BF16 (expect to add) AMX_TILE (expect to add) AMX_INT8 (expect to add)

reference: intel architecture instruction set extensions programming reference

[whyb]

AVX512ER and AVX512PF are used in the OpenVINO framework. AVX512_4FMAPS and AVX512_4VNNIW are used by the OpenVINO framework. AVX512_VPOPCNTDQ is used by the OpenVINO framework.

https://github.com/openvinotoolkit/openvino/blob/releases/2024/0/src/core/reference/src/op/jit_generator.cpp#L44