__lasx_xvffintl_d_w
LASX intrinsic · group: Floating Point Conversion
Synopsis
__m256d __lasx_xvffintl_d_w (__m256i a)
#include <lasxintrin.h>
Instruction: xvffintl.d.w xr, xr
CPU Flags: LASX
Description
Convert signed 32-bit integer elements in lower part of every 128-bit vector of a to double precision floating point numbers.
Operation
for (int i = 0; i < 2; i++) {
dst.fp64[i] = (f64)(s32)a.word[i]; // rounding mode is not expressed in C
}
for (int i = 2; i < 4; i++) {
dst.fp64[i] = (f64)(s32)a.word[i + 2]; // rounding mode is not expressed in C
}
// Expands to:
if (0) {
dst.fp64[0] = (f64)((s32)a.word[0]);
dst.fp64[1] = (f64)((s32)a.word[1]);
dst.fp64[2] = (f64)((s32)a.word[4]);
dst.fp64[3] = (f64)((s32)a.word[5]);
}
Tested on real machine.
Latency and Throughput
| CPU | µarch | Latency | Throughput (IPC) |
|---|---|---|---|
| 3C5000 | LA464 | 5 | 1 |
| 3A6000 | LA664 | 5 | 2 |
| 3C6000 | LA664 | 5 | 2 |