Lines Matching defs:__b
44 _mm_add_ss(__m128 __a, __m128 __b)
46 __a[0] += __b[0];
51 _mm_add_ps(__m128 __a, __m128 __b)
53 return __a + __b;
57 _mm_sub_ss(__m128 __a, __m128 __b)
59 __a[0] -= __b[0];
64 _mm_sub_ps(__m128 __a, __m128 __b)
66 return __a - __b;
70 _mm_mul_ss(__m128 __a, __m128 __b)
72 __a[0] *= __b[0];
77 _mm_mul_ps(__m128 __a, __m128 __b)
79 return __a * __b;
83 _mm_div_ss(__m128 __a, __m128 __b)
85 __a[0] /= __b[0];
90 _mm_div_ps(__m128 __a, __m128 __b)
92 return __a / __b;
135 _mm_min_ss(__m128 __a, __m128 __b)
137 return __builtin_ia32_minss(__a, __b);
141 _mm_min_ps(__m128 __a, __m128 __b)
143 return __builtin_ia32_minps(__a, __b);
147 _mm_max_ss(__m128 __a, __m128 __b)
149 return __builtin_ia32_maxss(__a, __b);
153 _mm_max_ps(__m128 __a, __m128 __b)
155 return __builtin_ia32_maxps(__a, __b);
159 _mm_and_ps(__m128 __a, __m128 __b)
161 return (__m128)((__v4si)__a & (__v4si)__b);
165 _mm_andnot_ps(__m128 __a, __m128 __b)
167 return (__m128)(~(__v4si)__a & (__v4si)__b);
171 _mm_or_ps(__m128 __a, __m128 __b)
173 return (__m128)((__v4si)__a | (__v4si)__b);
177 _mm_xor_ps(__m128 __a, __m128 __b)
179 return (__m128)((__v4si)__a ^ (__v4si)__b);
183 _mm_cmpeq_ss(__m128 __a, __m128 __b)
185 return (__m128)__builtin_ia32_cmpss(__a, __b, 0);
189 _mm_cmpeq_ps(__m128 __a, __m128 __b)
191 return (__m128)__builtin_ia32_cmpps(__a, __b, 0);
195 _mm_cmplt_ss(__m128 __a, __m128 __b)
197 return (__m128)__builtin_ia32_cmpss(__a, __b, 1);
201 _mm_cmplt_ps(__m128 __a, __m128 __b)
203 return (__m128)__builtin_ia32_cmpps(__a, __b, 1);
207 _mm_cmple_ss(__m128 __a, __m128 __b)
209 return (__m128)__builtin_ia32_cmpss(__a, __b, 2);
213 _mm_cmple_ps(__m128 __a, __m128 __b)
215 return (__m128)__builtin_ia32_cmpps(__a, __b, 2);
219 _mm_cmpgt_ss(__m128 __a, __m128 __b)
222 __builtin_ia32_cmpss(__b, __a, 1),
227 _mm_cmpgt_ps(__m128 __a, __m128 __b)
229 return (__m128)__builtin_ia32_cmpps(__b, __a, 1);
233 _mm_cmpge_ss(__m128 __a, __m128 __b)
236 __builtin_ia32_cmpss(__b, __a, 2),
241 _mm_cmpge_ps(__m128 __a, __m128 __b)
243 return (__m128)__builtin_ia32_cmpps(__b, __a, 2);
247 _mm_cmpneq_ss(__m128 __a, __m128 __b)
249 return (__m128)__builtin_ia32_cmpss(__a, __b, 4);
253 _mm_cmpneq_ps(__m128 __a, __m128 __b)
255 return (__m128)__builtin_ia32_cmpps(__a, __b, 4);
259 _mm_cmpnlt_ss(__m128 __a, __m128 __b)
261 return (__m128)__builtin_ia32_cmpss(__a, __b, 5);
265 _mm_cmpnlt_ps(__m128 __a, __m128 __b)
267 return (__m128)__builtin_ia32_cmpps(__a, __b, 5);
271 _mm_cmpnle_ss(__m128 __a, __m128 __b)
273 return (__m128)__builtin_ia32_cmpss(__a, __b, 6);
277 _mm_cmpnle_ps(__m128 __a, __m128 __b)
279 return (__m128)__builtin_ia32_cmpps(__a, __b, 6);
283 _mm_cmpngt_ss(__m128 __a, __m128 __b)
286 __builtin_ia32_cmpss(__b, __a, 5),
291 _mm_cmpngt_ps(__m128 __a, __m128 __b)
293 return (__m128)__builtin_ia32_cmpps(__b, __a, 5);
297 _mm_cmpnge_ss(__m128 __a, __m128 __b)
300 __builtin_ia32_cmpss(__b, __a, 6),
305 _mm_cmpnge_ps(__m128 __a, __m128 __b)
307 return (__m128)__builtin_ia32_cmpps(__b, __a, 6);
311 _mm_cmpord_ss(__m128 __a, __m128 __b)
313 return (__m128)__builtin_ia32_cmpss(__a, __b, 7);
317 _mm_cmpord_ps(__m128 __a, __m128 __b)
319 return (__m128)__builtin_ia32_cmpps(__a, __b, 7);
323 _mm_cmpunord_ss(__m128 __a, __m128 __b)
325 return (__m128)__builtin_ia32_cmpss(__a, __b, 3);
329 _mm_cmpunord_ps(__m128 __a, __m128 __b)
331 return (__m128)__builtin_ia32_cmpps(__a, __b, 3);
335 _mm_comieq_ss(__m128 __a, __m128 __b)
337 return __builtin_ia32_comieq(__a, __b);
341 _mm_comilt_ss(__m128 __a, __m128 __b)
343 return __builtin_ia32_comilt(__a, __b);
347 _mm_comile_ss(__m128 __a, __m128 __b)
349 return __builtin_ia32_comile(__a, __b);
353 _mm_comigt_ss(__m128 __a, __m128 __b)
355 return __builtin_ia32_comigt(__a, __b);
359 _mm_comige_ss(__m128 __a, __m128 __b)
361 return __builtin_ia32_comige(__a, __b);
365 _mm_comineq_ss(__m128 __a, __m128 __b)
367 return __builtin_ia32_comineq(__a, __b);
371 _mm_ucomieq_ss(__m128 __a, __m128 __b)
373 return __builtin_ia32_ucomieq(__a, __b);
377 _mm_ucomilt_ss(__m128 __a, __m128 __b)
379 return __builtin_ia32_ucomilt(__a, __b);
383 _mm_ucomile_ss(__m128 __a, __m128 __b)
385 return __builtin_ia32_ucomile(__a, __b);
389 _mm_ucomigt_ss(__m128 __a, __m128 __b)
391 return __builtin_ia32_ucomigt(__a, __b);
395 _mm_ucomige_ss(__m128 __a, __m128 __b)
397 return __builtin_ia32_ucomige(__a, __b);
401 _mm_ucomineq_ss(__m128 __a, __m128 __b)
403 return __builtin_ia32_ucomineq(__a, __b);
471 _mm_cvtsi32_ss(__m128 __a, int __b)
473 __a[0] = __b;
478 _mm_cvt_si2ss(__m128 __a, int __b)
480 return _mm_cvtsi32_ss(__a, __b);
486 _mm_cvtsi64_ss(__m128 __a, long long __b)
488 __a[0] = __b;
495 _mm_cvtpi32_ps(__m128 __a, __m64 __b)
497 return __builtin_ia32_cvtpi2ps(__a, (__v2si)__b);
501 _mm_cvt_pi2ps(__m128 __a, __m64 __b)
503 return _mm_cvtpi32_ps(__a, __b);
519 __mm_loadh_pi_v2f32 __b = ((struct __mm_loadh_pi_struct*)__p)->__u;
520 __m128 __bb = __builtin_shufflevector(__b, __b, 0, 1, 0, 1);
531 __mm_loadl_pi_v2f32 __b = ((struct __mm_loadl_pi_struct*)__p)->__u;
532 __m128 __bb = __builtin_shufflevector(__b, __b, 0, 1, 0, 1);
701 __v4hi __b = (__v4hi)__a;
702 return (unsigned short)__b[__n & 3];
708 __v4hi __b = (__v4hi)__a;
709 __b[__n & 3] = __d;
710 return (__m64)__b;
714 _mm_max_pi16(__m64 __a, __m64 __b)
716 return (__m64)__builtin_ia32_pmaxsw((__v4hi)__a, (__v4hi)__b);
720 _mm_max_pu8(__m64 __a, __m64 __b)
722 return (__m64)__builtin_ia32_pmaxub((__v8qi)__a, (__v8qi)__b);
726 _mm_min_pi16(__m64 __a, __m64 __b)
728 return (__m64)__builtin_ia32_pminsw((__v4hi)__a, (__v4hi)__b);
732 _mm_min_pu8(__m64 __a, __m64 __b)
734 return (__m64)__builtin_ia32_pminub((__v8qi)__a, (__v8qi)__b);
744 _mm_mulhi_pu16(__m64 __a, __m64 __b)
746 return (__m64)__builtin_ia32_pmulhuw((__v4hi)__a, (__v4hi)__b);
760 _mm_avg_pu8(__m64 __a, __m64 __b)
762 return (__m64)__builtin_ia32_pavgb((__v8qi)__a, (__v8qi)__b);
766 _mm_avg_pu16(__m64 __a, __m64 __b)
768 return (__m64)__builtin_ia32_pavgw((__v4hi)__a, (__v4hi)__b);
772 _mm_sad_pu8(__m64 __a, __m64 __b)
774 return (__m64)__builtin_ia32_psadbw((__v8qi)__a, (__v8qi)__b);
791 __m128 __b = (b); \
792 (__m128)__builtin_shufflevector((__v4sf)__a, (__v4sf)__b, \
798 _mm_unpackhi_ps(__m128 __a, __m128 __b)
800 return __builtin_shufflevector(__a, __b, 2, 6, 3, 7);
804 _mm_unpacklo_ps(__m128 __a, __m128 __b)
806 return __builtin_shufflevector(__a, __b, 0, 4, 1, 5);
810 _mm_move_ss(__m128 __a, __m128 __b)
812 return __builtin_shufflevector(__a, __b, 4, 1, 2, 3);
816 _mm_movehl_ps(__m128 __a, __m128 __b)
818 return __builtin_shufflevector(__a, __b, 6, 7, 2, 3);
822 _mm_movelh_ps(__m128 __a, __m128 __b)
824 return __builtin_shufflevector(__a, __b, 0, 1, 4, 5);
830 __m64 __b, __c;
833 __b = _mm_setzero_si64();
834 __b = _mm_cmpgt_pi16(__b, __a);
835 __c = _mm_unpackhi_pi16(__a, __b);
839 __c = _mm_unpacklo_pi16(__a, __b);
848 __m64 __b, __c;
851 __b = _mm_setzero_si64();
852 __c = _mm_unpackhi_pi16(__a, __b);
856 __c = _mm_unpacklo_pi16(__a, __b);
865 __m64 __b;
867 __b = _mm_setzero_si64();
868 __b = _mm_cmpgt_pi8(__b, __a);
869 __b = _mm_unpacklo_pi8(__a, __b);
871 return _mm_cvtpi16_ps(__b);
877 __m64 __b;
879 __b = _mm_setzero_si64();
880 __b = _mm_unpacklo_pi8(__a, __b);
882 return _mm_cvtpi16_ps(__b);
886 _mm_cvtpi32x2_ps(__m64 __a, __m64 __b)
891 __c = _mm_cvtpi32_ps(__c, __b);
900 __m64 __b, __c;
902 __b = _mm_cvtps_pi32(__a);
906 return _mm_packs_pi32(__b, __c);
912 __m64 __b, __c;
914 __b = _mm_cvtps_pi16(__a);
917 return _mm_packs_pi16(__b, __c);