__lsx_vsrar_h

LSX intrinsic · group: Shift

Synopsis

__m128i __lsx_vsrar_h (__m128i a, __m128i b)
#include <lsxintrin.h>
Instruction: vsrar.h vr, vr, vr
CPU Flags: LSX

Description

Arithmetic right shift with rounding each signed 16-bit element in a by the corresponding element of b modulo 16, store the result in dst.

Operation

for (int i = 0; i < 8; i++) {
  if ((b.half[i] & 0xf) == 0) {
    dst.half[i] = a.half[i];
  } else {
    dst.half[i] = ((s16)a.half[i] >> (b.half[i] & 0xf)) +
                  (((s16)a.half[i] >> ((b.half[i] & 0xf) - 1)) & 0x1);
  }
}

Tested on real machine.

Latency and Throughput

CPU	µarch	Latency	Throughput (IPC)
3C5000	LA464	3	2
3A6000	LA664	3	2
3C6000	LA664	3	2
2K1000LA	LA264	2	1
2K3000	LA364E	2	1

---

All intrinsics in alphabetical order