__lsx_vfdiv_s
LSX intrinsic · group: Floating Point Computation
Synopsis
__m128 __lsx_vfdiv_s (__m128 a, __m128 b)
#include <lsxintrin.h>
Instruction: vfdiv.s vr, vr, vr
CPU Flags: LSX
Description
Divide corresponding single precision floating point elements in a by elements in b.
Operation
for (int i = 0; i < 4; i++) {
dst.fp32[i] = a.fp32[i] / b.fp32[i];
}
Tested on real machine.
Latency and Throughput
| CPU | µarch | Latency | Throughput (IPC) |
|---|---|---|---|
| 3C5000 | LA464 | 11, 19.5 | 0.13(1/7.5) |
| 3A6000 | LA664 | 11 | 0.18(1/5.5) |
| 3C6000 | LA664 | 11 | 0.22(1/4.5) |
| 2K1000LA | LA264 | 12, 28 | 0.04(1/28) |
| 2K3000 | LA364E | 12, 28 | 0.04(1/28) |