Skip to content

VEXPANDPD

Henk-Jan Lebbink edited this page Jun 5, 2018 · 14 revisions

VEXPANDPD — Load Sparse Packed Double-Precision Floating-Point Values from Dense Memory

Opcode/ Instruction Op / En 64/32 bit Mode Support CPUID Feature Flag Description
EVEX.128.66.0F38.W1 88 /r VEXPANDPD xmm1 {k1}{z}, xmm2/m128 A V/V AVX512VL AVX512F Expand packed double-precision floating-point values from xmm2/m128 to xmm1 using writemask k1.
EVEX.256.66.0F38.W1 88 /r VEXPANDPD ymm1 {k1}{z}, ymm2/m256 A V/V AVX512VL AVX512F Expand packed double-precision floating-point values from ymm2/m256 to ymm1 using writemask k1.
EVEX.512.66.0F38.W1 88 /r VEXPANDPD zmm1 {k1}{z}, zmm2/m512 A V/V AVX512F Expand packed double-precision floating-point values from zmm2/m512 to zmm1 using writemask k1.

Instruction Operand Encoding

Op/En Tuple Type Operand 1 Operand 2 Operand 3 Operand 4
A Tuple1 Scalar ModRM:reg (w) ModRM:r/m (r) NA NA

Description

Expand (load) up to 8/4/2, contiguous, double-precision floating-point values of the input vector in the source operand (the second operand) to sparse elements in the destination operand (the first operand) selected by the writemask k1.

The destination operand is a ZMM/YMM/XMM register, the source operand can be a ZMM/YMM/XMM register or a 512/256/128-bit memory location.

The input vector starts from the lowest element in the source operand. The writemask register k1 selects the destination elements (a partial vector or sparse elements if less than 8 elements) to be replaced by the ascending elements in the input vector. Destination elements not selected by the writemask k1 are either unmodified or zeroed, depending on EVEX.z.

EVEX.vvvv is reserved and must be 1111b otherwise instructions will #UD.

Note that the compressed displacement assumes a pre-scaling (N) corresponding to the size of one single element instead of the size of the full vector.

Operation

VEXPANDPD (EVEX encoded versions)

(KL, VL) = (2, 128), (4, 256), (8, 512)
k0
FOR j0 TO KL-1
    ij * 64
    IF k1[j] OR *no writemask*
        THEN 
            DEST[i+63:i] ← SRC[k+63:k];
            kk + 64
        ELSE 
            IF *merging-masking*
                            ; merging-masking
                THEN *DEST[i+63:i] remains unchanged*
                ELSE 
                            ; zeroing-masking
                    THEN DEST[i+63:i] ← 0
            FI
    FI;
ENDFOR
DEST[MAXVL-1:VL] ←0

Intel C/C++ Compiler Intrinsic Equivalent

VEXPANDPD __m512d _mm512_mask_expand_pd( __m512d s, __mmask8 k, __m512d a);
VEXPANDPD __m512d _mm512_maskz_expand_pd( __mmask8 k, __m512d a);
VEXPANDPD __m512d _mm512_mask_expandloadu_pd( __m512d s, __mmask8 k, void * a);
VEXPANDPD __m512d _mm512_maskz_expandloadu_pd( __mmask8 k, void * a);
VEXPANDPD __m256d _mm256_mask_expand_pd( __m256d s, __mmask8 k, __m256d a);
VEXPANDPD __m256d _mm256_maskz_expand_pd( __mmask8 k, __m256d a);
VEXPANDPD __m256d _mm256_mask_expandloadu_pd( __m256d s, __mmask8 k, void * a);
VEXPANDPD __m256d _mm256_maskz_expandloadu_pd( __mmask8 k, void * a);
VEXPANDPD __m128d _mm_mask_expand_pd( __m128d s, __mmask8 k, __m128d a);
VEXPANDPD __m128d _mm_maskz_expand_pd( __mmask8 k, __m128d a);
VEXPANDPD __m128d _mm_mask_expandloadu_pd( __m128d s, __mmask8 k, void * a);
VEXPANDPD __m128d _mm_maskz_expandloadu_pd( __mmask8 k, void * a);

SIMD Floating-Point Exceptions

None

Other Exceptions

See Exceptions Type E4.nb.

#UD If EVEX.vvvv != 1111B.


Source: Intel® Architecture Software Developer's Manual (May 2018)
Generated: 5-6-2018

Clone this wiki locally