maki/openairinterface5g
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(nr_codeword_scrambling): Use SIMD XOR for scrambling

Browse Source
This commit is contained in:
Romain Beurdouche
2025-05-13 14:08:33 +00:00
parent 0aa3dba5b7
commit 4d092495bf
1 changed files with 19 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files
openair1/PHY/NR_TRANSPORT/nr_scrambling.c
+19 -25
View File
@@ -33,33 +33,27 @@ void nr_codeword_scrambling(uint8_t *in,
{
const int roundedSz = (size + 31) / 32;
uint32_t *seq = gold_cache((n_RNTI << 15) + (q << 14) + Nid, roundedSz);
for (int i = 0; i < roundedSz; i++)
out[i] = ((uint32_t*)in)[i] ^ seq[i];
#if 0
//#elif __AVX2__
for (int i = 0; i < roundedSz; i++) {
__m256i c = ((__m256i*)in)[i];
uint32_t in32 = _mm256_movemask_epi8(_mm256_slli_epi16(c, 7));
out[i] = in32 ^ seq[i];
DEBUG_SCRAMBLING(LOG_D(PHY, "in[%d] %x => %x\n", i, in32, out[i]));
}
//#elif defined (__aarch64__)
const int8_t __attribute__ ((aligned (16))) ucShift[] = {0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7};
int8x16_t vshift = vld1q_s8(ucShift);
uint8x16_t *c = (uint8x16_t*)in;
uint8x16_t cshift;
uint32_t in32;
for (int i = 0; i < roundedSz; i++) {
cshift = vshlq_u8(*c++, vshift);
in32 = vaddv_u8(vget_low_u8(cshift));
in32 += (vaddv_u8(vget_high_u8(cshift)) << 8);
cshift = vshlq_u8(*c++, vshift);
in32 += (vaddv_u8(vget_low_u8(cshift)) << 16);
in32 += (vaddv_u8(vget_high_u8(cshift)) << 24);
out[i] = in32 ^ seq[i];
DEBUG_SCRAMBLING(LOG_D(PHY, "in[%d] %x => %x\n", i, in32, out[i]));
unsigned int i_32 = 0;
#if defined(__AVX512F__) && defined(__AVX512VL__)
for (; i_32 < ((roundedSz >> 3) << 3); i_32 += 8) {
__m256i in_256 = _mm256_load_epi32(&((uint32_t *)in)[i_32]);
__m256i seq_256 = _mm256_load_epi32(&seq[i_32]);
_mm256_storeu_epi32(&out[i_32], _mm256_xor_si256(in_256, seq_256));
}
#endif
#if defined(__aarch64__)
for (; i_32 < ((roundedSz >> 2) << 2); i_32 += 4) {
unsigned int i = i_32 >> 2;
uint32x4_t in_32x4 = vld1q_u32(&((uint32_t *)in)[i_32]);
uint32x4_t seq_32x4 = vld1q_u32(&seq[i_32]);
((uint32x4_t *)out)[i] = veorq_u32(in_32x4, seq_32x4);
}
#endif
for (; i_32 < roundedSz; i_32++) {
unsigned int i = i_32;
out[i] = ((uint32_t *)in)[i] ^ seq[i];
DEBUG_SCRAMBLING(LOG_D(PHY, "in[%d] %x => %x\n", i, ((uint32_t*)in)[i], out[i]));
}
}
void nr_codeword_unscrambling(int16_t* llr, uint32_t size, uint8_t q, uint32_t Nid, uint32_t n_RNTI)