x86-64.net

Instruction

VPSUBD

Subtracts packed doublewords, wrapping around on overflow.

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

VPSUBD subtracts the doublewords of its second source from those of its first and keeps the low 32 bits of each difference, so the result wraps around, the same for signed and unsigned numbers, and no flag records it: on Zen 5, vpsubd xmm4, xmm4, [mem] turned 7FFFFFFFh minus FFFFFFFFh into 80000000h. It is the VEX and EVEX form of PSUBD, with what those encodings add to every integer operation, as the VPADDD page describes: a separate destination, a memory operand at any address, clearing above the vector width, and in EVEX a write mask and a broadcast doubleword.

The order of the operands

Only the second source can come from memory or be broadcast, and subtraction does not commute. GCC 14 and Clang 23 at -O2 with -march=x86-64-v3 compile a[i] = b[i] - c[i] over ints to a load of b and VPSUBD with c from memory. With -march=x86-64-v4 both compile _mm512_sub_epi32(x, _mm512_set1_epi32(*p)) to vpsubd zmm0, zmm0, dword ptr [rdi]{1to16}, but *p - x first needs a VPBROADCASTD into a register. A constant is not subtracted at all: x - 5 becomes VPADDD of -5.

There is no vector negation: both compilers compute -b[i] as VPSUBD from a register zeroed with VPXOR. And a compare’s -1 per element can be subtracted to count: loops counting a[i] == k or a[i] > k subtract each VPCMPEQD or VPCMPGTD result from an accumulator.

On Zen 5

A register minus itself is zero, and AMD’s Zen 5 guide lists VPSUBB, VPSUBW, VPSUBD and VPSUBQ with the same register twice among the zeroing idioms, which take no cycle and do not wait for the old value (Section 2.9.2); the compilers zero with VPXOR or VXORPS instead. Otherwise the latency spreadsheet gives VPSUBD what it gives VPADDD: one cycle, four per cycle. Zen 5’s 10 recorded cases cover every form; vpsubd xmm12, xmm6, xmm6 left all 512 bits of ZMM12 zero.

Forms

Every encoding of VPSUBD that Intel XED knows, one row per XED instruction form (iform). How to read the notation.
InstructionEncodingRequires
VPSUBD xmm, xmm, m128 VPSUBD_XMMdq_XMMdq_MEMdqVEX.128.66.0F.WIG FA /rAVX
VPSUBD xmm, xmm, xmm VPSUBD_XMMdq_XMMdq_XMMdqVEX.128.66.0F.WIG FA /rAVX
VPSUBD ymm, ymm, m256 VPSUBD_YMMqq_YMMqq_MEMqqVEX.256.66.0F.WIG FA /rAVX2
VPSUBD ymm, ymm, ymm VPSUBD_YMMqq_YMMqq_YMMqqVEX.256.66.0F.WIG FA /rAVX2
VPSUBD zmm {k}{z}, zmm, zmm VPSUBD_ZMMu32_MASKmskw_ZMMu32_ZMMu32_AVX512EVEX.512.66.0F.W0 FA /rAVX10.1 or AVX512F
VPSUBD zmm {k}{z}, zmm, m512/m32bcst VPSUBD_ZMMu32_MASKmskw_ZMMu32_MEMu32_AVX512EVEX.512.66.0F.W0 FA /rAVX10.1 or AVX512F
VPSUBD xmm {k}{z}, xmm, xmm VPSUBD_XMMu32_MASKmskw_XMMu32_XMMu32_AVX512EVEX.128.66.0F.W0 FA /rAVX10.1 or AVX512F + AVX512VL
VPSUBD xmm {k}{z}, xmm, m128/m32bcst VPSUBD_XMMu32_MASKmskw_XMMu32_MEMu32_AVX512EVEX.128.66.0F.W0 FA /rAVX10.1 or AVX512F + AVX512VL
VPSUBD ymm {k}{z}, ymm, ymm VPSUBD_YMMu32_MASKmskw_YMMu32_YMMu32_AVX512EVEX.256.66.0F.W0 FA /rAVX10.1 or AVX512F + AVX512VL
VPSUBD ymm {k}{z}, ymm, m256/m32bcst VPSUBD_YMMu32_MASKmskw_YMMu32_MEMu32_AVX512EVEX.256.66.0F.W0 FA /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, m128
xmm {k}{z}, xmm, m128/m32bcst		xmm, xmm, xmm
xmm {k}{z}, xmm, xmm
ymm {k}{z}, ymm, ymm		ymm, ymm, m256
ymm {k}{z}, ymm, m256/m32bcst		ymm, ymm, ymmzmm {k}{z}, zmm, zmmno maskzmm {k}{z}, zmm, zmmwith a mask, zeroingzmm {k}{z}, zmm, m512/m32bcst
Sandy Bridge7 / 0.5p151 / 0.5p15
Haswell7 / 0.5p151 / 0.5p158 / 0.5p151 / 0.5p15
Broadwell6 / 0.5p151 / 0.5p157 / 0.5p151 / 0.5p15
Skylake7 / 0.5p0151 / 0.33p0158 / 0.5p0151 / 0.33p015
Skylake-SP, Ice Lake7 / 0.5p0151 / 0.33p0158 / 0.5p0151 / 0.33p0151 / 0.33p0151 / 0.33p0158 / 0.5p015
Alder Lake8 / 0.33p0151 / 0.33p0159 / 0.33p0151 / 0.33p015
Sapphire Rapids8 / 0.33p0151 / 0.33p0159 / 0.33p0151 / 0.33p0151 / 0.5p051 / 0.5p059 / 0.5p05
Lunar Lake7 / 2V0 V1 V2 V31 / 0.25V0 V1 V2 V38 / 2.33V0 V1 V2 V31 / 0.25V0 V1 V2 V3
Piledriver7 / 1.5FPMAL2 / 0.5FPMAL FPU2 FPU3
Zen8 / 0.5FPU0 FPU1 FPU31 / 0.33FPU0 FPU1 FPU38 / 0.67FPU0 FPU1 FPU31 / 0.67FPU0 FPU1 FPU3
Zen 28 / 0.33FPU0 FPU1 FPU31 / 0.33FPU0 FPU1 FPU38 / 0.33FPU0 FPU1 FPU31 / 0.33FPU0 FPU1 FPU3
Zen 38 / 0.5FP0 FP1 FP2 FP31 / 0.25FP0 FP1 FP2 FP38 / 0.5FP0 FP1 FP2 FP31 / 0.25FP0 FP1 FP2 FP3
Zen 48 / 0.5FP0 FP1 FP2 FP31 / 0.25FP0 FP1 FP2 FP38 / 0.5FP0 FP1 FP2 FP31 / 0.25FP0 FP1 FP2 FP31 / 0.25FP0 FP1 FP2 FP31 / 0.5FP0 FP1 FP2 FP38 / 0.5FP0 FP1 FP2 FP3
Jaguar6 / 1FPU0 FPU1 VALU0 VALU11 / 0.5FPU0 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.

Element-wise: each doubleword of the result comes from the doubleword in the same place in each source.
VPSUBD: element by elementResult doubleword i comes from doubleword i of xmm2 and xmm3. Result doublewords 3 to 0: a3-b3, unchanged (bit 2 of k1 is 0), a1-b1, a0-b0.0123xmm2a1110k1a0-b0a1-b1a3-b3xmm10123xmm3or m128b

Every doubleword works alike: 4, 8 and 16 doublewords in the xmm, ymm and zmm forms. The EVEX forms write under a mask: where its bit is 0, the element keeps its old value, or zeroes it with {z}. With m32bcst, one doubleword from memory stands for every element of the last source.

Result doubleword i comes from doubleword i of xmm2 and xmm3. Result doublewords 3 to 0: a3-b3, unchanged (bit 2 of k1 is 0), a1-b1, a0-b0.

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] = zmm2.d[i] - src3.d[i]
    elif zeroing:
        zmm1.d[i] = 0
    # else: zmm1.d[i] keeps its value
# m32bcst: src3.d[i] is the m32 for every i
# 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
VPSUBD xmm, xmm, xmm/m128 VPSUBD ymm, ymm, ymm/m256
xmm, ymmexplicitwrite
xmm, ymmexplicitread
xmm/m128, ymm/m256explicitread
VPSUBD xmm {k}{z}, xmm, xmm/m128/m32bcst VPSUBD ymm {k}{z}, ymm, ymm/m256/m32bcst VPSUBD zmm {k}{z}, zmm, zmm/m512/m32bcst
xmm, ymm, zmmexplicitwrite (masked)
{k}explicitread (write mask)
xmm, ymm, zmmexplicitread
xmm/m128/m32bcst, ymm/m256/m32bcst, zmm/m512/m32bcstexplicitread

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-4
EVEX forms
AVX512-E4

Attributes

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

EVEX forms
write masking
EVEX memory forms
embedded broadcast (tuple Full); memory fault suppression

Intrinsics

The C intrinsics of Clang 23.1.2 that compile to VPSUBD, 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_sub_epi32 __m128i _mm_sub_epi32(__m128i __a, __m128i __b)

VPSUBD xmm, xmm, xmm VEX.128 -mavxwith AVX enabled

_mm_mask_sub_epi32 __m128i _mm_mask_sub_epi32(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B)

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

_mm_maskz_sub_epi32 __m128i _mm_maskz_sub_epi32(__mmask8 __U, __m128i __A, __m128i __B)

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

_mm256_sub_epi32 __m256i _mm256_sub_epi32(__m256i __a, __m256i __b)

VPSUBD ymm, ymm, ymm VEX.256 -mavx2

_mm256_mask_sub_epi32 __m256i _mm256_mask_sub_epi32(__m256i __W, __mmask8 __U, __m256i __A, __m256i __B)

VPSUBD ymm {k}, ymm, ymm EVEX.256 -mavx512vl

_mm256_maskz_sub_epi32 __m256i _mm256_maskz_sub_epi32(__mmask8 __U, __m256i __A, __m256i __B)

VPSUBD ymm {k}{z}, ymm, ymm EVEX.256 -mavx512vl

_mm512_sub_epi32 __m512i _mm512_sub_epi32(__m512i __A, __m512i __B)

VPSUBD zmm, zmm, zmm EVEX.512 -mavx512f

_mm512_mask_sub_epi32 __m512i _mm512_mask_sub_epi32(__m512i __W, __mmask16 __U, __m512i __A, __m512i __B)

VPSUBD zmm {k}, zmm, zmm EVEX.512 -mavx512f

_mm512_maskz_sub_epi32 __m512i _mm512_maskz_sub_epi32(__mmask16 __U, __m512i __A, __m512i __B)

VPSUBD zmm {k}{z}, zmm, zmm 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 4-479 (PSUBB/PSUBW/PSUBD)
  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.