x86-64.net

Instruction

VPERMI2W

Selects words from two tables with an index vector that the result then overwrites.

Extensions
Forms
6 (6 EVEX)
Data
Intel XED v2026.08.23

Forms

Every encoding of VPERMI2W that Intel XED knows, one row per XED instruction form (iform). How to read the notation.
InstructionEncodingRequires
VPERMI2W xmm {k}{z}, xmm, xmm VPERMI2W_XMMu16_MASKmskw_XMMu16_XMMu16_AVX512EVEX.128.66.0F38.W1 75 /rAVX10.1 or AVX512BW + AVX512VL
VPERMI2W xmm {k}{z}, xmm, m128 VPERMI2W_XMMu16_MASKmskw_XMMu16_MEMu16_AVX512EVEX.128.66.0F38.W1 75 /rAVX10.1 or AVX512BW + AVX512VL
VPERMI2W ymm {k}{z}, ymm, ymm VPERMI2W_YMMu16_MASKmskw_YMMu16_YMMu16_AVX512EVEX.256.66.0F38.W1 75 /rAVX10.1 or AVX512BW + AVX512VL
VPERMI2W ymm {k}{z}, ymm, m256 VPERMI2W_YMMu16_MASKmskw_YMMu16_MEMu16_AVX512EVEX.256.66.0F38.W1 75 /rAVX10.1 or AVX512BW + AVX512VL
VPERMI2W zmm {k}{z}, zmm, zmm VPERMI2W_ZMMu16_MASKmskw_ZMMu16_ZMMu16_AVX512EVEX.512.66.0F38.W1 75 /rAVX10.1 or AVX512BW
VPERMI2W zmm {k}{z}, zmm, m512 VPERMI2W_ZMMu16_MASKmskw_ZMMu16_MEMu16_AVX512EVEX.512.66.0F38.W1 75 /rAVX10.1 or AVX512BW

Intrinsics

The C intrinsics of Clang 23.1.2 that compile to VPERMI2W, each with the form it compiles to and the Clang options it needs. The forms come from compiling every intrinsic and decoding the result with Intel XED; all intrinsics A–Z explains the method.

IntrinsicCompiles to
_mm_permutex2var_epi16 __m128i _mm_permutex2var_epi16(__m128i __A, __m128i __I, __m128i __B)

VPERMI2W xmm, xmm, xmm EVEX.128 -mavx512vl -mavx512bwwhen __I is already in the destination register

_mm_permutex2var_pbh __m128bh _mm_permutex2var_pbh(__m128bh __A, __m128i __I, __m128bh __B)

VPERMI2W xmm, xmm, xmm EVEX.128 -mavx10.2when __A comes from memory

_mm_permutex2var_ph __m128h _mm_permutex2var_ph(__m128h __A, __m128i __I, __m128h __B)

VPERMI2W xmm, xmm, xmm EVEX.128 -mavx512fp16 -mavx512vlwhen __A comes from memory

_mm_mask2_permutex2var_epi16 __m128i _mm_mask2_permutex2var_epi16(__m128i __A, __m128i __I, __mmask8 __U, __m128i __B)

VPERMI2W xmm {k}, xmm, xmm EVEX.128 -mavx512vl -mavx512bw

_mm_maskz_permutex2var_epi16 __m128i _mm_maskz_permutex2var_epi16(__mmask8 __U, __m128i __A, __m128i __I, __m128i __B)

VPERMI2W xmm {k}{z}, xmm, xmm EVEX.128 -mavx512vl -mavx512bwwhen __I is already in the destination register

_mm256_permutex2var_epi16 __m256i _mm256_permutex2var_epi16(__m256i __A, __m256i __I, __m256i __B)

VPERMI2W ymm, ymm, ymm EVEX.256 -mavx512vl -mavx512bwwhen __I is already in the destination register

_mm256_permutex2var_pbh __m256bh _mm256_permutex2var_pbh(__m256bh __A, __m256i __I, __m256bh __B)

VPERMI2W ymm, ymm, ymm EVEX.256 -mavx10.2when __A comes from memory

_mm256_permutex2var_ph __m256h _mm256_permutex2var_ph(__m256h __A, __m256i __I, __m256h __B)

VPERMI2W ymm, ymm, ymm EVEX.256 -mavx512fp16 -mavx512vlwhen __A comes from memory

_mm256_mask2_permutex2var_epi16 __m256i _mm256_mask2_permutex2var_epi16(__m256i __A, __m256i __I, __mmask16 __U, __m256i __B)

VPERMI2W ymm {k}, ymm, ymm EVEX.256 -mavx512vl -mavx512bw

_mm256_maskz_permutex2var_epi16 __m256i _mm256_maskz_permutex2var_epi16(__mmask16 __U, __m256i __A, __m256i __I, __m256i __B)

VPERMI2W ymm {k}{z}, ymm, ymm EVEX.256 -mavx512vl -mavx512bwwhen __I is already in the destination register

_mm512_permutex2var_epi16 __m512i _mm512_permutex2var_epi16(__m512i __A, __m512i __I, __m512i __B)

VPERMI2W zmm, zmm, zmm EVEX.512 -mavx512bwwhen __I is already in the destination register

_mm512_permutex2var_pbh __m512bh _mm512_permutex2var_pbh(__m512bh __A, __m512i __I, __m512bh __B)

VPERMI2W zmm, zmm, zmm EVEX.512 -mavx10.2when __A comes from memory

_mm512_permutex2var_ph __m512h _mm512_permutex2var_ph(__m512h __A, __m512i __I, __m512h __B)

VPERMI2W zmm, zmm, zmm EVEX.512 -mavx512fp16when __A comes from memory

_mm512_mask2_permutex2var_epi16 __m512i _mm512_mask2_permutex2var_epi16(__m512i __A, __m512i __I, __mmask32 __U, __m512i __B)

VPERMI2W zmm {k}, zmm, zmm EVEX.512 -mavx512bw

_mm512_maskz_permutex2var_epi16 __m512i _mm512_maskz_permutex2var_epi16(__mmask32 __U, __m512i __A, __m512i __I, __m512i __B)

VPERMI2W zmm {k}{z}, zmm, zmm EVEX.512 -mavx512bwwhen __I is already in the destination register

CPUID and processors

Each form belongs to an XED ISA set. A form can be used when the processor reports every CPUID bit of one of its ISA set's alternatives.

CPUID requirements and processors of the XED ISA sets above. Alternatives are separated by “or”; processors follow XED’s chip model, which covers AMD processors up to Zen 2.
ISA setCPUIDProcessors in XED
AVX512BW_128

AVX10.1
CPUID.(EAX=07H,ECX=1):EDX[19] AVX10
CPUID.(EAX=24H,ECX=0):EBX[7:0] ≥ 1 AVX10 version

or

AVX512BW + AVX512VL
CPUID.(EAX=07H,ECX=0):EBX[30] AVX512BW
CPUID.(EAX=07H,ECX=0):EBX[31] AVX512VL

Cannon Lake, Ice Lake (client), Tiger Lake, Nova Lake, Skylake-SP, Cascade Lake, Cooper Lake, Ice Lake (server), Sapphire Rapids, Emerald Rapids, Granite Rapids, Diamond Rapids
AVX512BW_256

AVX10.1
CPUID.(EAX=07H,ECX=1):EDX[19] AVX10
CPUID.(EAX=24H,ECX=0):EBX[7:0] ≥ 1 AVX10 version

or

AVX512BW + AVX512VL
CPUID.(EAX=07H,ECX=0):EBX[30] AVX512BW
CPUID.(EAX=07H,ECX=0):EBX[31] AVX512VL

Cannon Lake, Ice Lake (client), Tiger Lake, Nova Lake, Skylake-SP, Cascade Lake, Cooper Lake, Ice Lake (server), Sapphire Rapids, Emerald Rapids, Granite Rapids, Diamond Rapids
AVX512BW_512

AVX10.1
CPUID.(EAX=07H,ECX=1):EDX[19] AVX10
CPUID.(EAX=24H,ECX=0):EBX[7:0] ≥ 1 AVX10 version

or

AVX512BW
CPUID.(EAX=07H,ECX=0):EBX[30] AVX512BW

Cannon Lake, Ice Lake (client), Tiger Lake, Nova Lake, Skylake-SP, Cascade Lake, Cooper Lake, Ice Lake (server), Sapphire Rapids, Emerald Rapids, Granite Rapids, Diamond Rapids

More AVX10.1 instructions

All 686 AVX10.1 instructions

Sources

  1. Intel XED v2026.08.23 (commit 0bcb6237345c): forms, encodings, ISA sets, CPUID bits and chip model.
  2. Clang's x86 intrinsic headers, LLVM 23.1.2 (commit 85ac56026243): the intrinsics, their signatures and doc comments, and the Clang of that release, which compiled each of them.
  3. Intel Advanced Vector Extensions 10.2 Architecture Specification, revision 6.0 (361050-006, January 2026)

The tables are derived from Intel XED, Copyright Intel Corporation, licensed under the Apache License 2.0; x86-64.net converted and reformatted them. The intrinsics and their signatures come from Clang, part of the LLVM Project, licensed under the Apache License 2.0 with LLVM Exceptions. The text is our own.