x86-64.net

Instruction

VPMOVSXWD

Sign-extends packed words to doublewords.

Extensions
Forms
10 (4 VEX, 6 EVEX)
Data
Intel XED v2026.08.23

VPMOVSXWD is the VEX and EVEX form of SSE4.1’s PMOVSXWD, whose page covers the sign extension of words to doublewords and the SSE2 sequences it replaces. The source is half the result: 4 words for an xmm destination (AVX), 8 for ymm (AVX2) and 16 for zmm (AVX512F; the EVEX xmm and ymm forms also need AVX512VL, or AVX10.1), from memory at any alignment. On Zen 5, vpmovsxwd ymm8, xmm15 turned word 3, 8000h, into FFFF8000h and placed words 4 to 7 in the upper 128-bit lane, clearing bits 511:256 of ZMM8, and {evex} vpmovsxwd xmm1, xmm2 turned FFFFh into FFFFFFFFh and 7FFFh into 00007FFFh. The EVEX forms add a write mask per doubleword.

Where compilers widen

For int16_t values copied into int32_t, in our builds at -O2 with -march=x86-64-v3 or x86-64-v4, Clang 23 widens eight words at a time straight from memory, while GCC 14 loads a full vector and widens its halves, moving the upper one down with VEXTRACTI128, or VEXTRACTI64X4 on zmm registers for x86-64-v4. Other loops show when widening is worth it:

  • Converted to double, four words per VCVTDQ2PD are widened in an xmm register by Clang, vpmovsxwd xmm0, qword ptr [mem], and taken from a widened ymm or zmm register by GCC.
  • Multiplied into int32_t, c[i] = a[i] * b[i] becomes VPMOVSXWD of both factors and VPMULLD on zmm registers in GCC for x86-64-v4. Clang instead zero-extends, as VPMOVZXWD’s page explains.
  • Summed as products into an int32_t total, a dot product, the words are not widened at all: both compilers multiply and add pairs with VPMADDWD.

AMD’s latency spreadsheet for Zen 5 gives the xmm and ymm forms a latency of 2 cycles and the zmm form 4, two per cycle each. The way back is VPACKSSDW, which saturates, or AVX-512’s VPMOVSDW and VPMOVDW, which keep the order.

In the optimization guide

What the optimization guide says about VPMOVSXWD, a line per claim. Each links to the claim in the guide, with its sources and where they differ.

Crestmont and Gracemont

  • Vectors and AVX2: The 256-bit forms that widen or insert data loaded from memory take two dependent micro-ops.

Forms

Every encoding of VPMOVSXWD that Intel XED knows, one row per XED instruction form (iform). How to read the notation.
InstructionEncodingRequires
VPMOVSXWD xmm, xmm VPMOVSXWD_XMMdq_XMMqVEX.128.66.0F38.WIG 23 /rAVX
VPMOVSXWD xmm, m64 VPMOVSXWD_XMMdq_MEMqVEX.128.66.0F38.WIG 23 /rAVX
VPMOVSXWD ymm, xmm VPMOVSXWD_YMMqq_XMMdqVEX.256.66.0F38.WIG 23 /rAVX2
VPMOVSXWD ymm, m128 VPMOVSXWD_YMMqq_MEMdqVEX.256.66.0F38.WIG 23 /rAVX2
VPMOVSXWD zmm {k}{z}, ymm VPMOVSXWD_ZMMi32_MASKmskw_YMMi16_AVX512EVEX.512.66.0F38.WIG 23 /rAVX10.1 or AVX512F
VPMOVSXWD zmm {k}{z}, m256 VPMOVSXWD_ZMMi32_MASKmskw_MEMi16_AVX512EVEX.512.66.0F38.WIG 23 /rAVX10.1 or AVX512F
VPMOVSXWD xmm {k}{z}, xmm VPMOVSXWD_XMMi32_MASKmskw_XMMi16_AVX512EVEX.128.66.0F38.WIG 23 /rAVX10.1 or AVX512F + AVX512VL
VPMOVSXWD xmm {k}{z}, m64 VPMOVSXWD_XMMi32_MASKmskw_MEMi16_AVX512EVEX.128.66.0F38.WIG 23 /rAVX10.1 or AVX512F + AVX512VL
VPMOVSXWD ymm {k}{z}, xmm VPMOVSXWD_YMMi32_MASKmskw_XMMi16_AVX512EVEX.256.66.0F38.WIG 23 /rAVX10.1 or AVX512F + AVX512VL
VPMOVSXWD ymm {k}{z}, m128 VPMOVSXWD_YMMi32_MASKmskw_MEMi16_AVX512EVEX.256.66.0F38.WIG 23 /rAVX10.1 or AVX512F + AVX512VL

Latency and throughput

What the scheduling models of LLVM 23.1.2 give for each form, not measurements: latency / reciprocal throughput in cycles, the ports under them. A row is a model, or models that agree; a dash, a form that its processors lack. Memory forms: latency from the address, ports without the load and store units. Which processors each model covers.
Modelxmm, xmm
xmm {k}{z}, xmm		xmm, m64ymm, xmm
ymm {k}{z}, xmm		ymm, m128zmm {k}{z}, ymmzmm {k}{z}, m256
ymm {k}{z}, m128		xmm {k}{z}, m64
Sandy Bridge1 / 0.5p157 / 0.5p15
Haswell, Skylake1 / 1p56 / 1p53 / 1p59 / 1p5
Broadwell1 / 1p56 / 1p53 / 1p58 / 1p5
Skylake-SP1 / 1p56 / 1p53 / 1p59 / 1p53 / 1p510 / 1p57 / 1p5
Ice Lake1 / 0.5p157 / 0.5p153 / 1p59 / 1p53 / 1p510 / 1p57 / 0.5p15
Alder Lake1 / 0.5p158 / 0.5p153 / 1p511 / 1p5
Sapphire Rapids1 / 0.5p158 / 0.5p153 / 1p511 / 1p53 / 1p511 / 1p58 / 0.5p15
Lunar Lake1 / 0.5V2 V38 / 0.5V2 V33 / 0.5V2 V310 / 2.33V2 V3
Piledriver2 / 1FPU1 FPXBR7 / 1.5FPU1 FPXBR
Zen1 / 0.5FPU1 FPU28 / 0.5FPU1 FPU21 / 2FPU1 FPU28 / 2FPU1 FPU2
Zen 21 / 0.5FPU1 FPU28 / 0.5FPU1 FPU24 / 0.5FPU1 FPU28 / 0.5FPU1 FPU2
Zen 31 / 0.5FP1 FP28 / 0.5FP1 FP24 / 1.5FP1 FP211 / 1.5FP1 FP2
Zen 42 / 0.5FP1 FP28 / 0.5FP1 FP22 / 0.5FP1 FP211 / 1.5FP1 FP24 / 1FP1 FP211 / 1.5FP1 FP28 / 0.5FP1 FP2
Jaguar1 / 0.5FPU0 FPU1 VALU0 VALU16 / 1FPU0 FPU1 VALU0 VALU1

Data flow

Where the elements of the sources go, element 0 on the right; results that depend on a control value are drawn for an example. How to read.

Widen: each word becomes a doubleword, sign-extended, in the same order.
VPMOVSXWD: each element widenedResult doublewords 15 to 0: a15, unchanged (bit 14 of k1 is 0), a13, a12, a11, a10, a9, a8, a7, a6, a5, a4, a3, a2, a1, a0.0123456789101112131415ymm2or m256a1111111111111110k1a0a1a2a3a4a5a6a7a8a9a10a11a12a13a15zmm1

The grey upper half of each doubleword is the extension: copies of the sign bit. Shown for the zmm form; the xmm and ymm forms work the same on 4 and 8 words. The EVEX forms write under a mask: where its bit is 0, the element keeps its old value, or zeroes it with {z}.

Result doublewords 15 to 0: a15, unchanged (bit 14 of k1 is 0), a13, a12, a11, a10, a9, a8, a7, a6, a5, a4, a3, a2, a1, a0.

Operation

What the forms do, in pseudocode that our tests run against results recorded on real processors. How to read.

for i in range(VL // 32):
    if k1[i]:                   # 1 without a write mask
        zmm1.d[i] = src2.sw[i]  # sign-extended
    elif zeroing:
        zmm1.d[i] = 0
    # else: zmm1.d[i] keeps its value
# zmm1 is zeroed from bit VL up

Run on AMD Zen 5 (EPYC 9645): 10 cases.

Operands

What each form reads and writes, as Intel XED records it: the operands of its syntax, in order, and the implicit ones that the syntax does not name. How to read.
OperandKindAction
VPMOVSXWD xmm, xmm/m64 VPMOVSXWD ymm, xmm/m128
xmm, ymmexplicitwrite
xmm/m64/128explicitread
VPMOVSXWD xmm {k}{z}, xmm/m64 VPMOVSXWD ymm {k}{z}, xmm/m128 VPMOVSXWD zmm {k}{z}, ymm/m256
xmm, ymm, zmmexplicitwrite (masked)
{k}explicitread (write mask)
xmm/m64, xmm/m128, ymm/m256explicitread

Flags

None: XED records no flag that the forms test or change.

Exceptions

The exception class is XED's name for the set of exception conditions that a form follows. How to read.

VEX forms
avx-type-5
EVEX forms
AVX512-E5

Attributes

What else XED records about the forms. How to read.

EVEX forms
write masking
EVEX memory forms
memory fault suppression

Intrinsics

The C intrinsics of Clang 23.1.2 that compile to VPMOVSXWD, 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_cvtepi16_epi32 __m128i _mm_cvtepi16_epi32(__m128i __V)

VPMOVSXWD xmm, xmm VEX.128 -mavxwith AVX enabled

_mm_mask_cvtepi16_epi32 __m128i _mm_mask_cvtepi16_epi32(__m128i __W, __mmask8 __U, __m128i __A)

VPMOVSXWD xmm {k}, xmm EVEX.128 -mavx512vl

_mm_maskz_cvtepi16_epi32 __m128i _mm_maskz_cvtepi16_epi32(__mmask8 __U, __m128i __A)

VPMOVSXWD xmm {k}{z}, xmm EVEX.128 -mavx512vl

_mm256_cvtepi16_epi32 __m256i _mm256_cvtepi16_epi32(__m128i __V)

VPMOVSXWD ymm, xmm VEX.256 -mavx2

_mm256_mask_cvtepi16_epi32 __m256i _mm256_mask_cvtepi16_epi32(__m256i __W, __mmask8 __U, __m128i __A)

VPMOVSXWD ymm {k}, xmm EVEX.256 -mavx512vl

_mm256_maskz_cvtepi16_epi32 __m256i _mm256_maskz_cvtepi16_epi32(__mmask8 __U, __m128i __A)

VPMOVSXWD ymm {k}{z}, xmm EVEX.256 -mavx512vl

_mm512_cvtepi16_epi32 __m512i _mm512_cvtepi16_epi32(__m256i __A)

VPMOVSXWD zmm, ymm EVEX.512 -mavx512f

_mm512_mask_cvtepi16_epi32 __m512i _mm512_mask_cvtepi16_epi32(__m512i __W, __mmask16 __U, __m256i __A)

VPMOVSXWD zmm {k}, ymm EVEX.512 -mavx512f

_mm512_maskz_cvtepi16_epi32 __m512i _mm512_maskz_cvtepi16_epi32(__mmask16 __U, __m256i __A)

VPMOVSXWD zmm {k}{z}, ymm EVEX.512 -mavx512f

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, whose AMD processors end with Zen 2, and the CPU data of this site for Zen 3 to Zen 6.
ISA setCPUIDProcessors
AVX

AVX
CPUID.01H:ECX[28] AVX

Sandy Bridge, Ivy Bridge, Haswell, Broadwell, Skylake, Comet Lake, Cannon Lake, Ice Lake (client), Tiger Lake, Alder Lake, Arrow Lake, Panther Lake, Nova Lake, Skylake-SP, Cascade Lake, Cooper Lake, Ice Lake (server), Sapphire Rapids, Emerald Rapids, Granite Rapids, Diamond Rapids, Sierra Forest, Clearwater Forest, Knights Landing, Knights Mill, Bulldozer, Piledriver, Zen, Zen+, Zen 2, Zen 3, Zen 4, Zen 5, Zen 6
AVX2

AVX2
CPUID.(EAX=07H,ECX=0):EBX[5] AVX2

Haswell, Broadwell, Skylake, Comet Lake, Cannon Lake, Ice Lake (client), Tiger Lake, Alder Lake, Arrow Lake, Panther Lake, Nova Lake, Skylake-SP, Cascade Lake, Cooper Lake, Ice Lake (server), Sapphire Rapids, Emerald Rapids, Granite Rapids, Diamond Rapids, Sierra Forest, Clearwater Forest, Knights Landing, Knights Mill, Zen, Zen+, Zen 2, Zen 3, Zen 4, Zen 5, Zen 6
AVX512F_512

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

or

AVX512F
CPUID.(EAX=07H,ECX=0):EBX[16] AVX512F

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, Knights Landing, Knights Mill, Zen 4, Zen 5, Zen 6
AVX512F_128

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

or

AVX512F + AVX512VL
CPUID.(EAX=07H,ECX=0):EBX[16] AVX512F
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, Zen 4, Zen 5, Zen 6
AVX512F_256
Same as AVX512F_128Same as AVX512F_128

More AVX instructions

All 253 AVX instructions

More AVX2 instructions

All 138 AVX2 instructions

More AVX10.1 instructions

All 686 AVX10.1 instructions

Sources

  1. Intel XED v2026.08.23 (commit 0bcb6237345c): forms, encodings, operands, flags, exception classes, attributes, ISA sets, CPUID bits and chip model.
  2. The CPU data of x86-64.net: Zen 3 to Zen 6, from GCC’s -march feature sets and AMD’s Processor Programming References, which the page of each processor cites.
  3. 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.
  4. LLVM 23.1.2's x86 scheduling models (commit 85ac56026243): the latency, reciprocal throughput and ports of each form, as llvm-mca reports them.
  5. Intel 64 and IA-32 Architectures Software Developer's Manual, Volume 2 (325383-093, September 2026): Volume 2B, page 4-352 (PMOVSX)
  6. Intel 64 and IA-32 Architectures Software Developer's Manual, Volume 1 (253665-093, September 2026)
  7. Intel Advanced Vector Extensions 10.2 Architecture Specification, revision 7.0 (361050-007, April 2026)
  8. Software Optimization Guide for the AMD Zen5 Microarchitecture (58455, rev. 1.00, August 2024), Wayback Machine copy of AMD's ZIP

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 and the timings from LLVM's scheduling models, both part of the LLVM Project, licensed under the Apache License 2.0 with LLVM Exceptions. The text is our own.