Lines Matching refs:__A
147 _mm_store_ps (float *__P, __m128 __A)
149 vec_st((__v4sf)__A, 0, (__v4sf*)__P);
154 _mm_storeu_ps (float *__P, __m128 __A)
156 *(__m128_u *)__P = __A;
161 _mm_storer_ps (float *__P, __m128 __A)
168 __tmp = (__m128) vec_perm (__A, __A, permute_vector);
175 _mm_store1_ps (float *__P, __m128 __A)
177 __v4sf __va = vec_splat((__v4sf)__A, 0);
182 _mm_store_ps1 (float *__P, __m128 __A)
184 _mm_store1_ps (__P, __A);
196 _mm_move_ss (__m128 __A, __m128 __B)
200 return (vec_sel ((__v4sf)__A, (__v4sf)__B, mask));
212 _mm_store_ss (float *__P, __m128 __A)
214 *__P = ((__v4sf)__A)[0];
222 _mm_add_ss (__m128 __A, __m128 __B)
231 a = vec_splat (__A, 0);
235 float elements from __A. */
236 return (vec_sel (__A, c, mask));
238 __A[0] = __A[0] + __B[0];
239 return (__A);
244 _mm_sub_ss (__m128 __A, __m128 __B)
253 a = vec_splat (__A, 0);
257 float elements from __A. */
258 return (vec_sel (__A, c, mask));
260 __A[0] = __A[0] - __B[0];
261 return (__A);
266 _mm_mul_ss (__m128 __A, __m128 __B)
275 a = vec_splat (__A, 0);
279 float elements from __A. */
280 return (vec_sel (__A, c, mask));
282 __A[0] = __A[0] * __B[0];
283 return (__A);
288 _mm_div_ss (__m128 __A, __m128 __B)
297 a = vec_splat (__A, 0);
301 float elements from __A. */
302 return (vec_sel (__A, c, mask));
304 __A[0] = __A[0] / __B[0];
305 return (__A);
310 _mm_sqrt_ss (__m128 __A)
318 a = vec_splat (__A, 0);
321 * float elements from __A. */
322 return (vec_sel (__A, c, mask));
327 _mm_add_ps (__m128 __A, __m128 __B)
329 return (__m128) ((__v4sf)__A + (__v4sf)__B);
333 _mm_sub_ps (__m128 __A, __m128 __B)
335 return (__m128) ((__v4sf)__A - (__v4sf)__B);
339 _mm_mul_ps (__m128 __A, __m128 __B)
341 return (__m128) ((__v4sf)__A * (__v4sf)__B);
345 _mm_div_ps (__m128 __A, __m128 __B)
347 return (__m128) ((__v4sf)__A / (__v4sf)__B);
351 _mm_sqrt_ps (__m128 __A)
353 return (vec_sqrt ((__v4sf)__A));
357 _mm_rcp_ps (__m128 __A)
359 return (vec_re ((__v4sf)__A));
363 _mm_rsqrt_ps (__m128 __A)
365 return (vec_rsqrte (__A));
369 _mm_rcp_ss (__m128 __A)
377 a = vec_splat (__A, 0);
380 * float elements from __A. */
381 return (vec_sel (__A, c, mask));
385 _mm_rsqrt_ss (__m128 __A)
393 a = vec_splat (__A, 0);
396 * float elements from __A. */
397 return (vec_sel (__A, c, mask));
401 _mm_min_ss (__m128 __A, __m128 __B)
409 a = vec_splat ((__v4sf)__A, 0);
413 * float elements from __A. */
414 return (vec_sel ((__v4sf)__A, c, mask));
418 _mm_max_ss (__m128 __A, __m128 __B)
426 a = vec_splat (__A, 0);
430 * float elements from __A. */
431 return (vec_sel ((__v4sf)__A, c, mask));
435 _mm_min_ps (__m128 __A, __m128 __B)
437 __vector __bool int m = vec_cmpgt ((__v4sf) __B, (__v4sf) __A);
438 return vec_sel (__B, __A, m);
442 _mm_max_ps (__m128 __A, __m128 __B)
444 __vector __bool int m = vec_cmpgt ((__v4sf) __A, (__v4sf) __B);
445 return vec_sel (__B, __A, m);
450 _mm_and_ps (__m128 __A, __m128 __B)
452 return ((__m128)vec_and ((__v4sf)__A, (__v4sf)__B));
453 // return __builtin_ia32_andps (__A, __B);
457 _mm_andnot_ps (__m128 __A, __m128 __B)
459 return ((__m128)vec_andc ((__v4sf)__B, (__v4sf)__A));
463 _mm_or_ps (__m128 __A, __m128 __B)
465 return ((__m128)vec_or ((__v4sf)__A, (__v4sf)__B));
469 _mm_xor_ps (__m128 __A, __m128 __B)
471 return ((__m128)vec_xor ((__v4sf)__A, (__v4sf)__B));
478 _mm_cmpeq_ps (__m128 __A, __m128 __B)
480 return ((__m128)vec_cmpeq ((__v4sf)__A,(__v4sf) __B));
484 _mm_cmplt_ps (__m128 __A, __m128 __B)
486 return ((__m128)vec_cmplt ((__v4sf)__A, (__v4sf)__B));
490 _mm_cmple_ps (__m128 __A, __m128 __B)
492 return ((__m128)vec_cmple ((__v4sf)__A, (__v4sf)__B));
496 _mm_cmpgt_ps (__m128 __A, __m128 __B)
498 return ((__m128)vec_cmpgt ((__v4sf)__A, (__v4sf)__B));
502 _mm_cmpge_ps (__m128 __A, __m128 __B)
504 return ((__m128)vec_cmpge ((__v4sf)__A, (__v4sf)__B));
508 _mm_cmpneq_ps (__m128 __A, __m128 __B)
510 __v4sf temp = (__v4sf ) vec_cmpeq ((__v4sf) __A, (__v4sf)__B);
515 _mm_cmpnlt_ps (__m128 __A, __m128 __B)
517 return ((__m128)vec_cmpge ((__v4sf)__A, (__v4sf)__B));
521 _mm_cmpnle_ps (__m128 __A, __m128 __B)
523 return ((__m128)vec_cmpgt ((__v4sf)__A, (__v4sf)__B));
527 _mm_cmpngt_ps (__m128 __A, __m128 __B)
529 return ((__m128)vec_cmple ((__v4sf)__A, (__v4sf)__B));
533 _mm_cmpnge_ps (__m128 __A, __m128 __B)
535 return ((__m128)vec_cmplt ((__v4sf)__A, (__v4sf)__B));
539 _mm_cmpord_ps (__m128 __A, __m128 __B)
546 a = (__vector unsigned int) vec_abs ((__v4sf)__A);
554 _mm_cmpunord_ps (__m128 __A, __m128 __B)
561 a = (__vector unsigned int) vec_abs ((__v4sf)__A);
572 _mm_cmpeq_ss (__m128 __A, __m128 __B)
581 a = vec_splat ((__v4sf) __A, 0);
585 * float elements from __A. */
586 return ((__m128)vec_sel ((__v4sf)__A, c, mask));
590 _mm_cmplt_ss (__m128 __A, __m128 __B)
599 a = vec_splat ((__v4sf) __A, 0);
603 * float elements from __A. */
604 return ((__m128)vec_sel ((__v4sf)__A, c, mask));
608 _mm_cmple_ss (__m128 __A, __m128 __B)
617 a = vec_splat ((__v4sf) __A, 0);
621 * float elements from __A. */
622 return ((__m128)vec_sel ((__v4sf)__A, c, mask));
626 _mm_cmpgt_ss (__m128 __A, __m128 __B)
635 a = vec_splat ((__v4sf) __A, 0);
639 * float elements from __A. */
640 return ((__m128)vec_sel ((__v4sf)__A, c, mask));
644 _mm_cmpge_ss (__m128 __A, __m128 __B)
653 a = vec_splat ((__v4sf) __A, 0);
657 * float elements from __A. */
658 return ((__m128)vec_sel ((__v4sf)__A, c, mask));
662 _mm_cmpneq_ss (__m128 __A, __m128 __B)
671 a = vec_splat ((__v4sf) __A, 0);
676 * float elements from __A. */
677 return ((__m128)vec_sel ((__v4sf)__A, c, mask));
681 _mm_cmpnlt_ss (__m128 __A, __m128 __B)
690 a = vec_splat ((__v4sf) __A, 0);
694 * float elements from __A. */
695 return ((__m128)vec_sel ((__v4sf)__A, c, mask));
699 _mm_cmpnle_ss (__m128 __A, __m128 __B)
708 a = vec_splat ((__v4sf) __A, 0);
712 * float elements from __A. */
713 return ((__m128)vec_sel ((__v4sf)__A, c, mask));
717 _mm_cmpngt_ss (__m128 __A, __m128 __B)
726 a = vec_splat ((__v4sf) __A, 0);
730 * float elements from __A. */
731 return ((__m128)vec_sel ((__v4sf)__A, c, mask));
735 _mm_cmpnge_ss (__m128 __A, __m128 __B)
744 a = vec_splat ((__v4sf) __A, 0);
748 * float elements from __A. */
749 return ((__m128)vec_sel ((__v4sf)__A, c, mask));
753 _mm_cmpord_ss (__m128 __A, __m128 __B)
762 a = (__vector unsigned int) vec_abs ((__v4sf)__A);
768 * float elements from __A. */
769 return ((__m128)vec_sel ((__v4sf)__A, (__v4sf)c, mask));
773 _mm_cmpunord_ss (__m128 __A, __m128 __B)
782 a = (__vector unsigned int) vec_abs ((__v4sf)__A);
788 * float elements from __A. */
789 return ((__m128)vec_sel ((__v4sf)__A, (__v4sf)c, mask));
795 _mm_comieq_ss (__m128 __A, __m128 __B)
797 return (__A[0] == __B[0]);
801 _mm_comilt_ss (__m128 __A, __m128 __B)
803 return (__A[0] < __B[0]);
807 _mm_comile_ss (__m128 __A, __m128 __B)
809 return (__A[0] <= __B[0]);
813 _mm_comigt_ss (__m128 __A, __m128 __B)
815 return (__A[0] > __B[0]);
819 _mm_comige_ss (__m128 __A, __m128 __B)
821 return (__A[0] >= __B[0]);
825 _mm_comineq_ss (__m128 __A, __m128 __B)
827 return (__A[0] != __B[0]);
839 _mm_ucomieq_ss (__m128 __A, __m128 __B)
841 return (__A[0] == __B[0]);
845 _mm_ucomilt_ss (__m128 __A, __m128 __B)
847 return (__A[0] < __B[0]);
851 _mm_ucomile_ss (__m128 __A, __m128 __B)
853 return (__A[0] <= __B[0]);
857 _mm_ucomigt_ss (__m128 __A, __m128 __B)
859 return (__A[0] > __B[0]);
863 _mm_ucomige_ss (__m128 __A, __m128 __B)
865 return (__A[0] >= __B[0]);
869 _mm_ucomineq_ss (__m128 __A, __m128 __B)
871 return (__A[0] != __B[0]);
875 _mm_cvtss_f32 (__m128 __A)
877 return ((__v4sf)__A)[0];
883 _mm_cvtss_si32 (__m128 __A)
895 : "+wa" (__A),
900 res = __builtin_rint(__A[0]);
906 _mm_cvt_ss2si (__m128 __A)
908 return _mm_cvtss_si32 (__A);
916 _mm_cvtss_si64 (__m128 __A)
928 : "+wa" (__A),
933 res = __builtin_llrint(__A[0]);
940 _mm_cvtss_si64x (__m128 __A)
942 return _mm_cvtss_si64 ((__v4sf) __A);
969 _mm_cvtps_pi32 (__m128 __A)
976 temp = (__v4sf) vec_splat ((__vector long long)__A, 0);
984 _mm_cvt_ps2pi (__m128 __A)
986 return _mm_cvtps_pi32 (__A);
991 _mm_cvttss_si32 (__m128 __A)
994 float temp = __A[0];
1000 _mm_cvtt_ss2si (__m128 __A)
1002 return _mm_cvttss_si32 (__A);
1007 _mm_cvttss_si64 (__m128 __A)
1010 float temp = __A[0];
1017 _mm_cvttss_si64x (__m128 __A)
1020 float temp = __A[0];
1028 _mm_cvttps_pi32 (__m128 __A)
1034 temp = (__v4sf) vec_splat ((__vector long long)__A, 0);
1041 _mm_cvtt_ps2pi (__m128 __A)
1043 return _mm_cvttps_pi32 (__A);
1048 _mm_cvtsi32_ss (__m128 __A, int __B)
1051 __A[0] = temp;
1053 return __A;
1057 _mm_cvt_si2ss (__m128 __A, int __B)
1059 return _mm_cvtsi32_ss (__A, __B);
1065 _mm_cvtsi64_ss (__m128 __A, long long __B)
1068 __A[0] = temp;
1070 return __A;
1075 _mm_cvtsi64x_ss (__m128 __A, long long __B)
1077 return _mm_cvtsi64_ss (__A, __B);
1083 _mm_cvtpi32_ps (__m128 __A, __m64 __B)
1093 ((__vector unsigned long long)__A) [1]});
1097 _mm_cvt_pi2ps (__m128 __A, __m64 __B)
1099 return _mm_cvtpi32_ps (__A, __B);
1104 _mm_cvtpi16_ps (__m64 __A)
1110 vs8 = (__vector signed short) (__vector unsigned long long) { __A, __A };
1119 _mm_cvtpu16_ps (__m64 __A)
1127 vs8 = (__vector unsigned short) (__vector unsigned long long) { __A, __A };
1141 _mm_cvtpi8_ps (__m64 __A)
1148 vc16 = (__vector signed char) (__vector unsigned long long) { __A, __A };
1159 _mm_cvtpu8_ps (__m64 __A)
1168 vc16 = (__vector unsigned char) (__vector unsigned long long) { __A, __A };
1185 _mm_cvtpi32x2_ps (__m64 __A, __m64 __B)
1190 vi4 = (__vector signed int) (__vector unsigned long long) { __A, __B };
1197 _mm_cvtps_pi16 (__m128 __A)
1203 rounded = vec_rint(__A);
1212 _mm_cvtps_pi8 (__m128 __A)
1220 rounded = vec_rint(__A);
1230 _mm_shuffle_ps (__m128 __A, __m128 __B, int const __mask)
1250 return vec_perm ((__v4sf) __A, (__v4sf)__B, (__vector unsigned char)t);
1255 _mm_unpackhi_ps (__m128 __A, __m128 __B)
1257 return (__m128) vec_vmrglw ((__v4sf) __A, (__v4sf)__B);
1262 _mm_unpacklo_ps (__m128 __A, __m128 __B)
1264 return (__m128) vec_vmrghw ((__v4sf) __A, (__v4sf)__B);
1270 _mm_loadh_pi (__m128 __A, __m64 const *__P)
1272 __vector unsigned long long __a = (__vector unsigned long long)__A;
1281 _mm_storeh_pi (__m64 *__P, __m128 __A)
1283 __vector unsigned long long __a = (__vector unsigned long long) __A;
1290 _mm_movehl_ps (__m128 __A, __m128 __B)
1293 (__vector unsigned long long)__A);
1298 _mm_movelh_ps (__m128 __A, __m128 __B)
1300 return (__m128) vec_mergeh ((__vector unsigned long long)__A,
1307 _mm_loadl_pi (__m128 __A, __m64 const *__P)
1309 __vector unsigned long long __a = (__vector unsigned long long)__A;
1318 _mm_storel_pi (__m64 *__P, __m128 __A)
1320 __vector unsigned long long __a = (__vector unsigned long long) __A;
1330 _mm_movemask_ps (__m128 __A)
1343 vec_vbpermq ((__vector unsigned char) __A,
1369 _mm_extract_pi16 (__m64 const __A, int const __N)
1376 return ((__A >> (shiftr * 16)) & 0xffff);
1380 _m_pextrw (__m64 const __A, int const __N)
1382 return _mm_extract_pi16 (__A, __N);
1388 _mm_insert_pi16 (__m64 const __A, int const __D, int const __N)
1393 __m64 result = (__A & (~mask)) | (shiftD & mask);
1399 _m_pinsrw (__m64 const __A, int const __D, int const __N)
1401 return _mm_insert_pi16 (__A, __D, __N);
1407 _mm_max_pi16 (__m64 __A, __m64 __B)
1413 a = (__vector signed short)vec_splats (__A);
1421 m1.as_m64 = __A;
1438 _m_pmaxsw (__m64 __A, __m64 __B)
1440 return _mm_max_pi16 (__A, __B);
1445 _mm_max_pu8 (__m64 __A, __m64 __B)
1451 a = (__vector unsigned char)vec_splats (__A);
1460 m1.as_m64 = __A;
1474 _m_pmaxub (__m64 __A, __m64 __B)
1476 return _mm_max_pu8 (__A, __B);
1481 _mm_min_pi16 (__m64 __A, __m64 __B)
1487 a = (__vector signed short)vec_splats (__A);
1495 m1.as_m64 = __A;
1512 _m_pminsw (__m64 __A, __m64 __B)
1514 return _mm_min_pi16 (__A, __B);
1519 _mm_min_pu8 (__m64 __A, __m64 __B)
1525 a = (__vector unsigned char)vec_splats (__A);
1534 m1.as_m64 = __A;
1548 _m_pminub (__m64 __A, __m64 __B)
1550 return _mm_min_pu8 (__A, __B);
1555 _mm_movemask_pi8 (__m64 __A)
1563 return __builtin_bpermd (p, __A);
1567 _m_pmovmskb (__m64 __A)
1569 return _mm_movemask_pi8 (__A);
1575 _mm_mulhi_pu16 (__m64 __A, __m64 __B)
1590 a = (__vector unsigned short)vec_splats (__A);
1601 _m_pmulhuw (__m64 __A, __m64 __B)
1603 return _mm_mulhi_pu16 (__A, __B);
1609 _mm_shuffle_pi16 (__m64 __A, int const __N)
1638 a = vec_splats (__A);
1644 _m_pshufw (__m64 __A, int const __N)
1646 return _mm_shuffle_pi16 (__A, __N);
1653 _mm_maskmove_si64 (__m64 __A, __m64 __N, char *__P)
1661 tmp = (tmp & (~mask)) | (__A & mask);
1666 _m_maskmovq (__m64 __A, __m64 __N, char *__P)
1668 _mm_maskmove_si64 (__A, __N, __P);
1673 _mm_avg_pu8 (__m64 __A, __m64 __B)
1677 a = (__vector unsigned char)vec_splats (__A);
1684 _m_pavgb (__m64 __A, __m64 __B)
1686 return _mm_avg_pu8 (__A, __B);
1691 _mm_avg_pu16 (__m64 __A, __m64 __B)
1695 a = (__vector unsigned short)vec_splats (__A);
1702 _m_pavgw (__m64 __A, __m64 __B)
1704 return _mm_avg_pu16 (__A, __B);
1711 _mm_sad_pu8 (__m64 __A, __m64 __B)
1720 a = (__vector unsigned char) (__vector unsigned long long) { 0UL, __A };
1736 _m_psadbw (__m64 __A, __m64 __B)
1738 return _mm_sad_pu8 (__A, __B);
1743 _mm_stream_pi (__m64 *__P, __m64 __A)
1752 *__P = __A;
1757 _mm_stream_ps (float *__P, __m128 __A)
1766 _mm_store_ps (__P, __A);