Intel® C++ Compiler 16.0 User and Reference Guide
Gathers 2/4 packed single-precision floating point values from memory referenced by the given base address, qword indices and scale. The corresponding Intel® AVX2 instruction is VGATHERQPS.
extern __m128 _mm_mask_i64gather_ps(float const * base, __m128i vindex, const int scale); |
extern __m128 _mm256_mask_i64gather_ps(float const * base, __m256i vindex, const int scale); |
base |
the base address used to reference the loaded FP elements. |
vindex |
the vector of qword indices used to reference the loaded FP elements. |
scale |
32-bit scale used to address the loaded FP elements. |
The intrinsics load 2/4 packed single-precision floating-point values from memory using qword indices and updates the destination operand. The intrinsic _mm_i64gather_ps() also sets the upper 64-bits of the result to '0'.
Below is the pseudo-code for the intrinsics:
_mm_i64gather_ps():
result[31:0] = mem[base+vindex[63:0]*scale]; result[63:32] = mem[base+vindex[127:64]*scale]; result[127:64] = 0;
_mm256_i64gather_ps():
result[31:0] = mem[base+vindex[63:0]*scale]; result[63:32] = mem[base+vindex[127:64]*scale]; result[95:64] = mem[base+vindex[191:128]*scale]; result[127:96] = mem[base+vindex[255:192]*scale];
A 128/256-bit vector with unconditionally gathered single-precision FP values.