MOVDDUP

MOVDDUP copies one double into both halves of an XMM register: the low double of a register, or eight bytes of memory. It is SSE3’s broadcast of a double and the double-precision side of MOVSLDUP, which repeats the even singles; VMOVDDUP does the same in each 128-bit lane of a wider register, and VBROADCASTSD fills a whole YMM or ZMM register.

The memory form reads only eight bytes, and they need not be aligned, unlike most legacy SSE memory operands; the result, like every legacy SSE result, keeps bits 128 and up of the register (see SSE and AVX). Zen 5’s recorded load and register cases both put the low double in both halves and left the upper bits as they were. No floating-point operation happens, so a NaN or a denormal is copied as it is.

SSE3 and the baseline#

SSE3 is not part of the first x86-64 level, so code built for the default target does without MOVDDUP: GCC 14 broadcasts a double with unpcklpd xmm0, xmm0 (UNPCKLPD) and Clang 23 with MOVLHPS, after a separate load when the value comes from memory. Both refuse _mm_movedup_pd and _mm_loaddup_pd without -msse3.

-march=x86-64-v2 includes SSE3. In our builds at -O2 with it, both compilers load and broadcast _mm_set1_pd(*p) and _mm_loaddup_pd(p) with one movddup xmm0, qword ptr [rdi], and Clang broadcasts a register with MOVDDUP as well, for example the scalar of a[i] = b[i] * s before its vector loop. GCC keeps unpcklpd xmm0, xmm0 for a register, even for _mm_movedup_pd; the result is the same.

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(2):
    j = 0
    xmm1.q[i] = src2.q[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_5

Intrinsics

The C intrinsics of Clang 23.1.2 that compile to MOVDDUP, 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
• MOVSHDUP
• MOVSLDUP