MOVSLDUP

MOVSLDUP copies the even singles of its source over the odd ones: the result is elements 0, 0, 2 and 2. It is the twin of MOVSHDUP, whose page covers what the two share: SSE3, a memory operand of 16 bytes that must be aligned, unlike MOVDDUP’s eight, bits 128 and up of the register kept (see SSE and AVX), and no floating-point operation, so NaNs and denormals pass through as bits. VMOVSLDUP is the VEX and EVEX form.

Zen 5’s recorded cases follow the pattern: movsldup xmm12, xmmword ptr [mem] put bytes 0 to 3 of memory in elements 0 and 1 and bytes 8 to 11 in elements 2 and 3, and movsldup xmm7, xmm11 repeated elements 0 and 2 of XMM11.

The real parts of complex floats#

An array of float complex keeps each real part in an even element and each imaginary part in the odd one after it, so MOVSLDUP spreads the real parts across their pairs and MOVSHDUP the imaginary ones. That is the first step of a complex product: in our builds at -O2 with -march=x86-64-v2 -fcx-limited-range, GCC 14 compiles r[i] = a[i] * b[i] on float complex values to MOVSLDUP and MOVSHDUP of a[i], each multiplied by b[i] or its swapped parts, and an ADDSUBPS to combine them. Without the option, C’s rules about infinities and NaNs make both GCC and Clang 23 keep the product scalar with a call to __mulsc3 for the rare cases.

As a plain shuffle, __builtin_shufflevector(v, v, 0, 0, 2, 2) is MOVSLDUP in both compilers once SSE3 is enabled, and shufps xmm0, xmm0, 160 for the default target.

Forms

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.

Operation

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

for i in range(4):
    j = i - i % 2
    xmm1.d[i] = src2.d[j]
# bits 128 and up of the register are unchanged

Operands

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.

All forms

SSE_TYPE_4

Attributes

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

Memory forms

aligned memory operand

Intrinsics

The C intrinsics of Clang 23.1.2 that compile to MOVSLDUP, 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.

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.

More SSE3 instructions

• ADDSUBPD
• ADDSUBPS
• FISTTP
• HADDPD
• HADDPS
• HSUBPD
• HSUBPS
• LDDQU
• MOVDDUP
• MOVSHDUP