__lsx_vfrecip_s
LSX intrinsic · group: Floating Point Computation
Synopsis
__m128 __lsx_vfrecip_s (__m128 a)
#include <lsxintrin.h>
Instruction: vfrecip.s vr, vr
CPU Flags: LSX
Description
Compute reciprocal of single precision floating point elements in a.
Operation
for (int i = 0; i < 4; i++) {
dst.fp32[i] = 1 / a.fp32[i];
}
Latency and Throughput
| CPU | µarch | Latency | Throughput (IPC) |
|---|---|---|---|
| 3C5000 | LA464 | 27 | 0.14(1/7) |
| 3A6000 | LA664 | 11 | 0.18(1/5.5) |
| 3C6000 | LA664 | 27 | 0.12(1/8.5) |
| 2K1000LA | LA264 | 28 | 0.04(1/28) |
| 2K3000 | LA364E | 28 | 0.04(1/28) |