minor changes

This commit is contained in:
Xavier Arteaga 2020-04-12 20:31:48 +02:00 committed by Xavier Arteaga
parent d76c77c188
commit fa776b1243
1 changed files with 40 additions and 39 deletions

View File

@ -26,7 +26,6 @@
#ifdef LV_HAVE_SSE #ifdef LV_HAVE_SSE
#include <immintrin.h> #include <immintrin.h>
#include <srslte/phy/common/sequence.h>
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */
/** /**
@ -129,8 +128,8 @@ static inline uint32_t sequence_gen_LTE_pr_memless_step_x2(uint32_t state)
} }
/** /**
* Static precomputed array x1 and x2 * Static precomputed x1 and x2 states after Nc shifts
* ---------------------------------- * -------------------------------------------------------
* *
* The pre-computation of the Pseudo-Random sequences is based in their linearity properties. * The pre-computation of the Pseudo-Random sequences is based in their linearity properties.
* *
@ -148,7 +147,7 @@ static uint32_t sequence_x1_init = 0;
static uint32_t sequence_x2_init[SEQUENCE_SEED_LEN] = {}; static uint32_t sequence_x2_init[SEQUENCE_SEED_LEN] = {};
/** /**
* C constructor, pre-computes X1 and X2 initial states and sequences * C constructor, pre-computes X1 and X2 initial states
*/ */
__attribute__((constructor)) __attribute__((unused)) static void srslte_lte_pr_pregen() __attribute__((constructor)) __attribute__((unused)) static void srslte_lte_pr_pregen()
{ {
@ -169,7 +168,7 @@ __attribute__((constructor)) __attribute__((unused)) static void srslte_lte_pr_p
} }
} }
static void sequence_gen_LTE_pr_memless(uint8_t* pr, uint32_t len, uint32_t seed) static void sequence_gen_LTE_pr(uint8_t* pr, uint32_t len, uint32_t seed)
{ {
int n = 0; int n = 0;
uint32_t x1 = sequence_x1_init; // X1 initial state is fix uint32_t x1 = sequence_x1_init; // X1 initial state is fix
@ -183,7 +182,7 @@ static void sequence_gen_LTE_pr_memless(uint8_t* pr, uint32_t len, uint32_t seed
} }
// Parallel stage // Parallel stage
if (len > SEQUENCE_PAR_BITS) { if (len >= SEQUENCE_PAR_BITS) {
for (; n < len - (SEQUENCE_PAR_BITS - 1); n += SEQUENCE_PAR_BITS) { for (; n < len - (SEQUENCE_PAR_BITS - 1); n += SEQUENCE_PAR_BITS) {
// XOR x1 and x2 // XOR x1 and x2
uint32_t c = (uint32_t)(x1 ^ x2); uint32_t c = (uint32_t)(x1 ^ x2);
@ -218,7 +217,7 @@ int srslte_sequence_set_LTE_pr(srslte_sequence_t* q, uint32_t len, uint32_t seed
return SRSLTE_ERROR; return SRSLTE_ERROR;
} }
sequence_gen_LTE_pr_memless(q->c, len, seed); sequence_gen_LTE_pr(q->c, len, seed);
return SRSLTE_SUCCESS; return SRSLTE_SUCCESS;
} }
@ -235,42 +234,44 @@ sequence_generate_signed(const uint8_t* c_unpacked, int8_t* c_char, int16_t* c_s
__m128i* sse_c_short = (__m128i*)c_short; __m128i* sse_c_short = (__m128i*)c_short;
float* sse_c_float = c_float; float* sse_c_float = c_float;
for (; i < ((int)len) - 15; i += 16) { if (len >= 16) {
// Get bit mask for (; i < len - 15; i += 16) {
__m128i m8 = _mm_cmpgt_epi8(_mm_load_si128(sse_c), _mm_set1_epi8(0)); // Get bit mask
sse_c++; __m128i m8 = _mm_cmpgt_epi8(_mm_load_si128(sse_c), _mm_set1_epi8(0));
sse_c++;
// Generate blend masks // Generate blend masks
__m128i m16_1 = _mm_unpacklo_epi8(m8, m8); __m128i m16_1 = _mm_unpacklo_epi8(m8, m8);
__m128i m16_2 = _mm_unpackhi_epi8(m8, m8); __m128i m16_2 = _mm_unpackhi_epi8(m8, m8);
__m128 m32_1 = (__m128)_mm_unpacklo_epi8(m16_1, m16_1); __m128 m32_1 = (__m128)_mm_unpacklo_epi8(m16_1, m16_1);
__m128 m32_2 = (__m128)_mm_unpackhi_epi8(m16_1, m16_1); __m128 m32_2 = (__m128)_mm_unpackhi_epi8(m16_1, m16_1);
__m128 m32_3 = (__m128)_mm_unpacklo_epi8(m16_2, m16_2); __m128 m32_3 = (__m128)_mm_unpacklo_epi8(m16_2, m16_2);
__m128 m32_4 = (__m128)_mm_unpackhi_epi8(m16_2, m16_2); __m128 m32_4 = (__m128)_mm_unpackhi_epi8(m16_2, m16_2);
// Generate int8 values // Generate int8 values
const __m128i bp = _mm_set1_epi8(+1); const __m128i bp = _mm_set1_epi8(+1);
const __m128i bn = _mm_set1_epi8(-1); const __m128i bn = _mm_set1_epi8(-1);
_mm_storeu_si128(sse_c_char, _mm_blendv_epi8(bp, bn, m8)); _mm_storeu_si128(sse_c_char, _mm_blendv_epi8(bp, bn, m8));
sse_c_char++; sse_c_char++;
// Generate int16 values // Generate int16 values
const __m128i sp = _mm_set1_epi16(+1); const __m128i sp = _mm_set1_epi16(+1);
const __m128i sn = _mm_set1_epi16(-1); const __m128i sn = _mm_set1_epi16(-1);
_mm_store_si128(sse_c_short++, _mm_blendv_epi8(sp, sn, m16_1)); _mm_store_si128(sse_c_short++, _mm_blendv_epi8(sp, sn, m16_1));
_mm_store_si128(sse_c_short++, _mm_blendv_epi8(sp, sn, m16_2)); _mm_store_si128(sse_c_short++, _mm_blendv_epi8(sp, sn, m16_2));
// Generate float values // Generate float values
const __m128 fp = _mm_set1_ps(+1); const __m128 fp = _mm_set1_ps(+1);
const __m128 fn = _mm_set1_ps(-1); const __m128 fn = _mm_set1_ps(-1);
_mm_store_ps(sse_c_float, _mm_blendv_ps(fp, fn, (__m128)m32_1)); _mm_store_ps(sse_c_float, _mm_blendv_ps(fp, fn, (__m128)m32_1));
sse_c_float += 4; sse_c_float += 4;
_mm_store_ps(sse_c_float, _mm_blendv_ps(fp, fn, (__m128)m32_2)); _mm_store_ps(sse_c_float, _mm_blendv_ps(fp, fn, (__m128)m32_2));
sse_c_float += 4; sse_c_float += 4;
_mm_store_ps(sse_c_float, _mm_blendv_ps(fp, fn, (__m128)m32_3)); _mm_store_ps(sse_c_float, _mm_blendv_ps(fp, fn, (__m128)m32_3));
sse_c_float += 4; sse_c_float += 4;
_mm_store_ps(sse_c_float, _mm_blendv_ps(fp, fn, (__m128)m32_4)); _mm_store_ps(sse_c_float, _mm_blendv_ps(fp, fn, (__m128)m32_4));
sse_c_float += 4; sse_c_float += 4;
}
} }
#endif /* LV_HAVE_SSE */ #endif /* LV_HAVE_SSE */