__lsx_vmulwev_h_bu_b
LSX intrinsic · group: Integer Computation
Synopsis
__m128i __lsx_vmulwev_h_bu_b (__m128i a, __m128i b)
#include <lsxintrin.h>
Instruction: vmulwev.h.bu.b vr, vr, vr
CPU Flags: LSX
Description
Multiply the even-positioned unsigned 8-bit elements in a and the corresponding signed 8-bit elements in b, extending both to 16-bit, and save the 16-bit result in dst.
Operation
for (int i = 0; i < 8; i++) {
dst.half[i] = (u16)(u8)a.byte[2 * i] * (s16)(s8)b.byte[2 * i];
}
// Expands to:
if (0) {
dst.half[0] = ((u16)((u8)a.byte[0])) * ((s16)((s8)b.byte[0]));
dst.half[1] = ((u16)((u8)a.byte[2])) * ((s16)((s8)b.byte[2]));
dst.half[2] = ((u16)((u8)a.byte[4])) * ((s16)((s8)b.byte[4]));
dst.half[3] = ((u16)((u8)a.byte[6])) * ((s16)((s8)b.byte[6]));
dst.half[4] = ((u16)((u8)a.byte[8])) * ((s16)((s8)b.byte[8]));
dst.half[5] = ((u16)((u8)a.byte[10])) * ((s16)((s8)b.byte[10]));
dst.half[6] = ((u16)((u8)a.byte[12])) * ((s16)((s8)b.byte[12]));
dst.half[7] = ((u16)((u8)a.byte[14])) * ((s16)((s8)b.byte[14]));
}
Tested on real machine.
Latency and Throughput
| CPU | µarch | Latency | Throughput (IPC) |
|---|---|---|---|
| 3C5000 | LA464 | 4 | 2 |
| 3A6000 | LA664 | 4 | 2 |
| 3C6000 | LA664 | 4 | 2 |
| 2K1000LA | LA264 | 4 | 1 |
| 2K3000 | LA364E | 4 | 1 |