x86 SIMD instruction listings

The x86 instruction set has several times been extended with SIMD (Single instruction, multiple data) instruction set extensions. These extensions, starting from the MMX instruction set extension introduced with Pentium MMX in 1997, typically define sets of wide registers and instructions that subdivide these registers into fixed-size lanes and perform a computation for each lane in parallel.

The main SIMD instruction set extensions that have been introduced for x86 are:

These instructions are, unless otherwise noted, available in the following forms:

• MMX: 64-bit vectors, operating on mm0..mm7 registers (aliased on top of the old x87 register file)
• SSE2: 128-bit vectors, operating on xmm0..xmm15 registers (xmm0..xmm7 in 32-bit mode)
• AVX: 128-bit vectors, operating on xmm0..xmm15 registers, with a new three-operand encoding enabled by the new VEX prefix. (AVX introduced 256-bit vector registers, but the full width of these vectors was in general not made available for integer SIMD instructions until AVX2.)
• AVX2: 256-bit vectors, operating on ymm0..ymm15 registers (extended versions of the xmm0..xmm15 registers)
• AVX-512: 512-bit vectors, operating on zmm0..zmm31 registers (zmm0..zmm15 are extended versions of the ymm0..ymm15 registers, while zmm16..zmm31 are new to AVX-512). AVX-512 also introduces opmasks, allowing the operation of most instructions to be masked on a per-lane basis by an opmask register (the lane width varies from one instruction to another). AVX-512 also adds broadcast functionality for many of its instructions - this is used with memory source arguments to replicate a single value to all lanes of a vector calculation. The tables below provide indications of whether opmasks and broadcasts are supported for each instruction, and if so, what lane-widths they are using.

For many of the instruction mnemonics, (V) is used to indicate that the instruction mnemonic exists in forms with and without a leading V - the form with the leading V is used for the VEX/EVEX-prefixed instruction variants introduced by AVX/AVX2/AVX-512, while the form without the leading V is used for legacy MMX/SSE encodings without VEX/EVEX-prefix.

For the instructions in the below table, the following considerations apply unless otherwise noted:

• Packed instructions are available at all vector lengths (128-bit for SSE2, 128/256-bit for AVX, 128/256/512-bit for AVX-512)
• FP32 variants of instructions are introduced as part of SSE. FP64 variants of instructions are introduced as part of SSE2.
• The AVX-512 variants of the FP32 and FP64 instructions are introduced as part of the AVX512F subset.
• For AVX-512 variants of the instructions, opmasks and broadcasts are available with a width of 32 bits for FP32 operations and 64 bits for FP64 operations. (Broadcasts are available for vector operations only.)

From SSE2 onwards, some data movement/bitwise instructions exist in three forms: an integer form, an FP32 form and an FP64 form. Such instructions are functionally identical, however some processors with SSE2 will implement integer, FP32 and FP64 execution units as three different execution clusters, where forwarding of results from one cluster to another may come with performance penalties and where such penalties can be minimzed by choosing instruction forms appropriately. (For example, there exists three forms of vector bitwise XOR instructions under SSE2 - PXOR, XORPS, and XORPD - these are intended for use on integer, FP32, and FP64 data, respectively.)

These instructions do not have any MMX forms, and do not support any encodings without a prefix. Most of these instructions have extended variants available in VEX-encoded and EVEX-encoded forms:

• The VEX-encoded forms are available under AVX/AVX2. Under AVX, they are available only with a vector length of 128 bits (VEX.L=0 enocding) - under AVX2, they are (with some exceptions noted with "L=0") also made available with a vector length of 256 bits.
• The EVEX-encoded forms are available under AVX-512 - the specific AVX-512 subset needed for each instruction is listed along with the instruction.

SSE SIMD instructions that do not fit into any of the preceding groups. Many of these instructions have AVX/AVX-512 extended forms - unless otherwise indicated (L=0 or footnotes) these extended forms support 128/256-bit operation under AVX and 128/256/512-bit operation under AVX-512.

AVX were first supported by Intel with Sandy Bridge and by AMD with Bulldozer.

Vector operations on 256 bit registers.

Half-precision floating-point conversion.

Introduced in Intel's Haswell microarchitecture and AMD's Excavator.

Expansion of most vector integer SSE and AVX instructions to 256 bits

Floating-point fused multiply-add instructions are introduced in x86 as two instruction set extensions, "FMA3" and "FMA4", both of which build on top of AVX to provide a set of scalar/vector instructions using the xmm/ymm/zmm vector registers. FMA3 defines a set of 3-operand fused-multiply-add instructions that take three input operands and writes its result back to the first of them. FMA4 defines a set of 4-operand fused-multiply-add instructions that take four input operands – a destination operand and three source operands.

FMA3 is supported on Intel CPUs starting with Haswell, on AMD CPUs starting with Piledriver, and on Zhaoxin CPUs starting with YongFeng. FMA4 was only supported on AMD Family 15h (Bulldozer) CPUs and has been abandoned from AMD Zen onwards. The FMA3/FMA4 extensions are not considered to be an intrinsic part of AVX or AVX2, although all Intel and AMD (but not Zhaoxin) processors that support AVX2 also support FMA3. FMA3 instructions (in EVEX-encoded form) are, however, AVX-512 foundation instructions.
The FMA3 and FMA4 instruction sets both define a set of 10 fused-multiply-add operations, all available in FP32 and FP64 variants. For each of these variants, FMA3 defines three operand orderings while FMA4 defines two.
FMA3 encoding
FMA3 instructions are encoded with the VEX or EVEX prefixes – on the form VEX.66.0F38 xy /r or EVEX.66.0F38 xy /r. The VEX.W/EVEX.W bit selects floating-point format (W=0 means FP32, W=1 means FP64). The opcode byte xy consists of two nibbles, where the top nibble x selects operand ordering (9='132', A='213', B='231') and the bottom nibble y (values 6..F) selects which one of the 10 fused-multiply-add operations to perform. (x and y outside the given ranges will result in something that is not an FMA3 instruction.)
At the assembly language level, the operand ordering is specified in the mnemonic of the instruction:

• vfmadd132sd xmm1,xmm2,xmm3 will perform xmm1 ← (xmm1*xmm3)+xmm2
• vfmadd213sd xmm1,xmm2,xmm3 will perform xmm1 ← (xmm2*xmm1)+xmm3
• vfmadd231sd xmm1,xmm2,xmm3 will perform xmm1 ← (xmm2*xmm3)+xmm1

For all FMA3 variants, the first two arguments must be xmm/ymm/zmm vector register arguments, while the last argument may be either a vector register or memory argument. Under AVX-512 and AVX10, the EVEX-encoded variants support EVEX-prefix-encoded broadcast, opmasks and rounding-controls.
The AVX512-FP16 extension, introduced in Sapphire Rapids, adds FP16 variants of the FMA3 instructions – these all take the form EVEX.66.MAP6.W0 xy /r with the opcode byte working in the same way as for the FP32/FP64 variants. The AVX10.2 extension, published in 2024, similarly adds BF16 variants of the packed (but not scalar) FMA3 instructions – these all take the form EVEX.NP.MAP6.W0 xy /r with the opcode byte again working similar to the FP32/FP64 variants. (For the FMA4 instructions, no FP16 or BF16 variants are defined.)
FMA4 encoding
FMA4 instructions are encoded with the VEX prefix, on the form VEX.66.0F3A xx /r ib (no EVEX encodings are defined). The opcode byte xx uses its bottom bit to select floating-point format (0=FP32, 1=FP64) and the remaining bits to select one of the 10 fused-multiply-add operations to perform.

For FMA4, operand ordering is controlled by the VEX.W bit. If VEX.W=0, then the third operand is the r/m operand specified by the instruction's ModR/M byte and the fourth operand is a register operand, specified by bits 7:4 of the ib (8-bit immediate) part of the instruction. If VEX.W=1, then these two operands are swapped. For example:

• vfmaddsd xmm1,xmm2,[mem],xmm3 will perform xmm1 ← (xmm2*[mem])+xmm3 and require a W=0 encoding.
• vfmaddsd xmm1,xmm2,xmm3,[mem] will perform xmm1 ← (xmm2*xmm3)+[mem] and require a W=1 encoding.
• vfmaddsd xmm1,xmm2,xmm3,xmm4 will perform xmm1 ← (xmm2*xmm3)+xmm4 and can be encoded with either W=0 or W=1.

Opcode table
The 10 fused-multiply-add operations and the 122 instruction variants they give rise to are given by the following table – with FMA4 instructions highlighted with * and yellow cell coloring, and FMA3 instructions not highlighted:

AVX-512, introduced in 2014, adds 512-bit wide vector registers (extending the 256-bit registers, which become the new registers' lower halves) and doubles their count to 32; the new registers are thus named zmm0 through zmm31. It adds eight mask registers, named k0 through k7, which may be used to restrict operations to specific parts of a vector register. Unlike previous instruction set extensions, AVX-512 is implemented in several groups; only the foundation ("AVX-512F") extension is mandatory. Most of the added instructions may also be used with the 256- and 128-bit registers.

Intel AMX adds eight new tile-registers, tmm0-tmm7, each holding a matrix, with a maximum capacity of 16 rows of 64 bytes per tile-register. It also adds a TILECFG register to configure the sizes of the actual matrices held in each of the eight tile-registers, and a set of instructions to perform matrix multiplications on these registers.

• x86 instruction listings
• List of discontinued x86 instructions
• ARM architecture family#64/32-bit architecture

• Intel Intrinsics Guide - searchable reference for Intel MMX/SSE/AVX/AVX512 SIMD intrinsics