__lsx_vfcvth_s_h
LSX intrinsic · group: Floating Point Conversion
Synopsis
__m128 __lsx_vfcvth_s_h (__m128i a)
#include <lsxintrin.h>
Instruction: vfcvth.s.h vr, vr
CPU Flags: LSX
Description
Convert half precision floating point elements in higher half of every 128-bit vector of a to single precision.
Operation
for (int i = 0; i < 4; i++) {
dst.fp32[i] = a.fp16[i + 4];
}
// Expands to:
if (0) {
dst.fp32[0] = a.fp16[4];
dst.fp32[1] = a.fp16[5];
dst.fp32[2] = a.fp16[6];
dst.fp32[3] = a.fp16[7];
}
Latency and Throughput
| CPU | µarch | Latency | Throughput (IPC) |
|---|---|---|---|
| 3C5000 | LA464 | 3 | 1 |
| 3A6000 | LA664 | 3 | 2 |
| 3C6000 | LA664 | 3 | 2 |
| 2K1000LA | LA264 | 3 | 0.5(1/2) |
| 2K3000 | LA364E | 3 | 0.5(1/2) |