����JFIFXX�����    $.' ",#(7),01444'9=82<.342  2!!22222222222222222222222222222222222222222222222222����"��4�� ���,�PG"Z_�4�˷����kjز�Z�,F+��_z�,�© �����zh6�٨�ic�fu���#ډb���_�N�?��wQ���5-�~�I���8����TK<5o�Iv-�����k�_U_�����~b�M��d����Ӝ�U�Hh��?]��E�w��Q���k�{��_}qFW7HTՑ��Y��F�?_�'ϔ��_�Ջt��=||I ��6�έ"�����D���/[�k�9���Y�8ds|\���Ҿp6�Ҵ���]��.����6�z<�v��@]�i%��$j��~�g��J>��no����pM[me�i$[����s�o�ᘨ�˸ nɜG-�ĨU�ycP�3.DB�li�;��hj���x7Z^�N�h������N3u{�:j�x�힞��#M&��jL P@_���� P��&��o8������9�����@Sz6�t7#O�ߋ �s}Yf�T���lmr����Z)'N��k�۞p����w\�Tȯ?�8`�O��i{wﭹW�[�r�� ��Q4F�׊���3m&L�=��h3����z~��#�\�l :�F,j@�� ʱ�wQT����8�"kJO���6�֚l����}���R�>ډK���]��y����&����p�}b��;N�1�m�r$�|��7�>e�@B�TM*-iH��g�D�)� E�m�|�ؘbҗ�a��Ҿ����t4���o���G��*oCN�rP���Q��@z,|?W[0�����:�n,jWiE��W��$~/�hp\��?��{(�0���+�Y8rΟ�+����>S-S����VN;�}�s?.����� w�9��˟<���Mq4�Wv'��{)0�1mB��V����W[�����8�/<� �%���wT^�5���b��)iM� pg�N�&ݝ��VO~�q���u���9� ����!��J27����$O-���! �:�%H��� ـ����y�ΠM=t{!S�� oK8������t<����è:a������[�����ա�H���~��w��Qz`�po�^ ����Q��n� �,uu�C�$ ^���,������8�#��:�6��e�|~���!�3�3.�\0��q��o�4`.|� ����y�Q�`~;�d�ׯ,��O�Zw�������`73�v�܋�<���Ȏ�� ـ4k��5�K�a�u�=9Yd��$>x�A�&�� j0� ���vF��� Y�|�y��� ~�6�@c��1vOp�Ig����4��l�OD���L����� R���c���j�_�uX6��3?nk��Wy�f;^*B� ��@�~a�`��Eu������+���6�L��.ü>��}y���}_�O�6�͐�:�YrG�X��kG�����l^w���~㒶sy��Iu�!� W ��X��N�7BV��O��!X�2����wvG�R�f�T#�����t�/?���%8�^�W�aT��G�cL�M���I��(J����1~�8�?aT ���]����AS�E��(��*E}� 2��#I/�׍qz��^t�̔���b�Yz4x���t�){ OH��+(E��A&�N�������XT��o��"�XC��'���)}�J�z�p� ��~5�}�^����+�6����w��c��Q�|Lp�d�H��}�(�.|����k��c4^�"�����Z?ȕ ��a<�L�!039C� �Eu�C�F�Ew�ç ;�n?�*o���B�8�bʝ���'#Rqf���M}7����]����s2tcS{�\icTx;�\��7K���P���ʇ Z O-��~��c>"��?�������P��E��O�8��@�8��G��Q�g�a�Վ���󁶠�䧘��_%#r�>�1�z�a��eb��qcPѵ��n���#L��� =��׀t� L�7�`��V���A{�C:�g���e@�w1 Xp3�c3�ġ����p��M"'-�@n4���fG��B3�DJ�8[Jo�ߐ���gK)ƛ��$���� ���8�3�����+���� �����6�ʻ���� ���S�kI�*KZlT _`���?��K����QK�d����B`�s}�>���`��*�>��,*@J�d�oF*����弝��O}�k��s��]��y�ߘ��c1G�V���<=�7��7����6�q�PT��tXԀ�!9*4�4Tހ3XΛex�46���Y��D ����� �BdemDa����\�_l,��G�/���֌7���Y�](�xTt^%�GE�����4�}bT���ڹ�����;Y)���B�Q��u��>J/J �⮶.�XԄ��j�ݳ�+E��d ��r�5�_D�1 ��o�� �B�x�΢�#���<��W�����8���R6�@g�M�.��� dr�D��>(otU��@x=��~v���2� ӣ�d�oBd��3�eO�6�㣷�����ݜ6��6Y��Qz`��S��{���\P�~z m5{J/L��1������<�e�ͅPu�b�]�ϔ���'������f�b� Zpw��c`"��i���BD@:)ִ�:�]��hv�E�w���T�l��P���"Ju�}��وV J��G6��. J/�Qgl߭�e�����@�z�Zev2u�)]կ�����7x���s�M�-<ɯ�c��r�v�����@��$�ޮ}lk���a���'����>x��O\�ZFu>�����ck#��&:��`�$�ai�>2Δ����l���oF[h��lE�ܺ�Πk:)���`�� $[6�����9�����kOw�\|���8}������ބ:��񶐕��I�A1/�=�2[�,�!��.}gN#�u����b��� ~��݊��}34q����d�E��Lc��$��"�[q�U�硬g^��%B �z���r�pJ�ru%v\h1Y�ne`ǥ:g���pQM~�^�Xi� ��`S�:V29.�P���V�?B�k�� AEvw%�_�9C�Q����wKekPؠ�\�;Io d�{ ߞo�c1eP����\� `����E=���@K<�Y���eڼ�J���w����{av�F�'�M�@/J��+9p���|]�����Iw &`��8���&M�hg��[�{��Xj��%��Ӓ�$��(����ʹN���<>�I���RY���K2�NPlL�ɀ)��&e����B+ь����( � �JTx���_?EZ� }@ 6�U���뙢ط�z��dWI�n` D����噥�[��uV��"�G&Ú����2g�}&m��?ċ�"����Om#��������� ��{�ON��"S�X��Ne��ysQ���@Fn��Vg���dX�~nj�]J�<�K]:��FW��b�������62�=��5f����JKw��bf�X�55��~J �%^����:�-�QIE��P��v�nZum� z � ~ə ���� ���ة����;�f��\v���g�8�1��f24;�V���ǔ�)����9���1\��c��v�/'Ƞ�w�������$�4�R-��t���� e�6�/�ġ �̕Ecy�J���u�B���<�W�ַ~�w[B1L۲�-JS΂�{���΃������A��20�c#��@ 0!1@AP"#2Q`$3V�%45a6�FRUq��� ����^7ׅ,$n�������+��F�`��2X'��0vM��p�L=������5��8������u�p~���.�`r�����\���O��,ư�0oS ��_�M�����l���4�kv\JSd���x���SW�<��Ae�IX����������$I���w�:S���y���›R��9�Q[���,�5�;�@]�%���u�@ *ro�lbI �� ��+���%m:�͇ZV�����u�̉����θau<�fc�.����{�4Ա� �Q����*�Sm��8\ujqs]{kN���)qO�y�_*dJ�b�7���yQqI&9�ԌK!�M}�R�;������S�T���1���i[U�ɵz�]��U)V�S6���3$K{�ߊ<�(� E]Զ[ǼENg�����'�\?#)Dkf��J���o��v���'�%ƞ�&K�u�!��b�35LX�Ϸ��63$K�a�;�9>,R��W��3�3� d�JeTYE.Mϧ��-�o�j3+y��y^�c�������VO�9NV\nd�1 ��!͕_)a�v;����թ�M�lWR1��)El��P;��yوÏ�u 3�k�5Pr6<�⒲l�!˞*��u־�n�!�l:����UNW ��%��Chx8vL'��X�@��*��)���̮��ˍ��� ���D-M�+J�U�kvK����+�x8��cY������?�Ԡ��~3mo��|�u@[XeY�C�\Kp�x8�oC�C�&����N�~3-H���� ��MX�s�u<`���~"WL��$8ξ��3���a�)|:@�m�\���^�`�@ҷ)�5p+��6���p�%i)P M���ngc�����#0Aruz���RL+xSS?���ʮ}()#�t��mˇ!��0}}y����<�e� �-ή�Ԩ��X������ MF���ԙ~l L.3���}�V뽺�v�����멬��Nl�)�2����^�Iq��a��M��qG��T�����c3#������3U�Ǎ���}��לS�|qa��ڃ�+���-��2�f����/��bz��ڐ�� �ݼ[2�ç����k�X�2�* �Z�d���J�G����M*9W���s{��w���T��x��y,�in�O�v��]���n����P�$�JB@=4�OTI�n��e�22a\����q�d���%�$��(���:���: /*�K[PR�fr\nڙdN���F�n�$�4�[�� U�zƶ����� �mʋ���,�ao�u 3�z� �x��Kn����\[��VFmbE;�_U��&V�Gg�]L�۪&#n%�$ɯ�dG���D�TI=�%+AB�Ru#��b4�1�»x�cs�YzڙJG��f��Il��d�eF'T� iA��T���uC�$����Y��H?����[!G`}���ͪ� �纤Hv\������j�Ex�K���!���OiƸ�Yj�+u-<���'q����uN�*�r\��+�]���<�wOZ.fp�ێ��,-*)V?j-kÊ#�`�r��dV����(�ݽBk�����G�ƛk�QmUڗe��Z���f}|����8�8��a���i��3'J�����~G_�^���d�8w������ R�`(�~�.��u���l�s+g�bv���W���lGc}��u���afE~1�Ue������Z�0�8�=e�� f@/�jqEKQQ�J��oN��J���W5~M>$6�Lt�;$ʳ{���^��6�{����v6���ķܰg�V�cnn �~z�x�«�,2�u�?cE+Ș�H؎�%�Za�)���X>uW�Tz�Nyo����s���FQƤ��$��*�&�LLXL)�1�" L��eO��ɟ�9=���:t��Z���c��Ž���Y?�ӭV�wv�~,Y��r�ۗ�|�y��GaF�����C�����.�+� ���v1���fήJ�����]�S��T��B��n5sW}y�$��~z�'�c ��8 ��� ,! �p��VN�S��N�N�q��y8z˱�A��4��*��'������2n<�s���^ǧ˭P�Jޮɏ�U�G�L�J�*#��<�V��t7�8����TĜ>��i}K%,���)[��z�21z ?�N�i�n1?T�I�R#��m-�����������������1����lA�`��fT5+��ܐ�c�q՝��ʐ��,���3�f2U�եmab��#ŠdQ�y>\��)�SLY����w#��.���ʑ�f��� ,"+�w�~�N�'�c�O�3F�������N<���)j��&��,-� �љ���֊�_�zS���TǦ����w�>��?�������n��U仆�V���e�����0���$�C�d���rP �m�׈e�Xm�Vu� �L��.�bֹ��� �[Դaզ���*��\y�8�Է:�Ez\�0�Kq�C b��̘��cө���Q��=0Y��s�N��S.���3.���O�o:���#���v7�[#߫ ��5�܎�L���Er4���9n��COWlG�^��0k�%<���ZB���aB_���������'=��{i�v�l�$�uC���mƎҝ{�c㱼�y]���W�i ��ߧc��m�H� m�"�"�����;Y�ߝ�Z�Ǔ�����:S#��|}�y�,/k�Ld� TA�(�AI$+I3��;Y*���Z��}|��ӧO��d�v��..#:n��f>�>���ȶI�TX��� 8��y����"d�R�|�)0���=���n4��6ⲑ�+��r<�O�܂~zh�z����7ܓ�HH�Ga롏���nCo�>������a ���~]���R���̲c?�6(�q�;5%� |�uj�~z8R=X��I�V=�|{v�Gj\gc��q����z�؋%M�ߍ����1y��#��@f^���^�>N�����#x#۹��6�Y~�?�dfPO��{��P�4��V��u1E1J �*|���%���JN��`eWu�zk M6���q t[�� ��g�G���v��WIG��u_ft����5�j�"�Y�:T��ɐ���*�;� e5���4����q$C��2d�}���� _S�L#m�Yp��O�.�C�;��c����Hi#֩%+) �Ӎ��ƲV���SYź��g |���tj��3�8���r|���V��1#;.SQ�A[���S������#���`n�+���$��$I �P\[�@�s��(�ED�z���P��])8�G#��0B��[ى��X�II�q<��9�~[Z멜�Z�⊔IWU&A>�P~�#��dp<�?����7���c��'~���5 ��+$���lx@�M�dm��n<=e�dyX��?{�|Aef ,|n3�<~z�ƃ�uۧ�����P��Y,�ӥQ�*g�#먙R�\���;T��i,��[9Qi歉����c>]9�� ��"�c��P�� �Md?٥��If�ت�u��k��/����F��9�c*9��Ǎ:�ØF���z�n*�@|I�ށ9����N3{'��[�'ͬ�Ҳ4��#}��!�V� Fu��,�,mTIk���v C�7v���B�6k�T9��1�*l� '~��ƞF��lU��'�M ����][ΩũJ_�{�i�I�n��$���L�� j��O�dx�����kza۪��#�E��Cl����x˘�o�����V���ɞ�ljr��)�/,�߬h�L��#��^��L�ф�,íMƁe�̩�NB�L�����iL����q�}��(��q��6IçJ$�W�E$��:������=#����(�K�B����zђ <��K(�N�۫K�w��^O{!����)�H���>x�������lx�?>Պ�+�>�W���,Ly!_�D���Ō�l���Q�!�[ �S����J��1��Ɛ�Y}��b,+�Lo�x�ɓ)����=�y�oh�@�꥟/��I��ѭ=��P�y9��� �ۍYӘ�e+�p�Jnϱ?V\SO%�(�t� ���=?MR�[Ș�����d�/ ��n�l��B�7j� ��!�;ӥ�/�[-���A�>�dN�sLj ��,ɪv��=1c�.SQ�O3�U���ƀ�ܽ�E����������̻��9G�ϷD�7(�}��Ävӌ\�y�_0[w ���<΍>����a_��[0+�L��F.�޺��f�>oN�T����q;���y\��bՃ��y�jH�<|q-eɏ�_?_9+P���Hp$�����[ux�K w�Mw��N�ی'$Y2�=��q���KB��P��~������Yul:�[<����F1�2�O���5=d����]Y�sw:���Ϯ���E��j,_Q��X��z`H1,#II ��d�wr��P˂@�ZJV����y$�\y�{}��^~���[:N����ߌ�U�������O��d�����ؾe��${p>G��3c���Ė�lʌ�� ת��[��`ϱ�-W����dg�I��ig2��� ��}s ��ؤ(%#sS@���~���3�X�nRG�~\jc3�v��ӍL��M[JB�T��s3}��j�Nʖ��W����;7��ç?=X�F=-�=����q�ߚ���#���='�c��7���ڑW�I(O+=:uxq�������������e2�zi+�kuG�R��������0�&e�n���iT^J����~\jy���p'dtG��s����O��3����9* �b#Ɋ�� p������[Bws�T�>d4�ۧs���nv�n���U���_�~,�v����ƜJ1��s�� �QIz��)�(lv8M���U=�;����56��G���s#�K���MP�=��LvyGd��}�VwWBF�'�à �?MH�U�g2�� ����!�p�7Q��j��ڴ����=��j�u��� Jn�A s���uM������e��Ɔ�Ҕ�!)'��8Ϣ�ٔ��ޝ(��Vp���צ֖d=�IC�J�Ǡ{q������kԭ�߸���i��@K����u�|�p=..�*+����x�����z[Aqġ#s2a�Ɗ���RR�)*HRsi�~�a &f��M��P����-K�L@��Z��Xy�'x�{}��Zm+���:�)�) IJ�-i�u���� ���ܒH��'�L(7�y�GӜq���� j��� 6ߌg1�g�o���,kر���tY�?W,���p���e���f�OQS��!K�۟cҒA�|ս�j�>��=⬒��˧L[�� �߿2JaB~R��u�:��Q�] �0H~���]�7��Ƽ�I���(}��cq '�ήET���q�?f�ab���ӥvr� �)o��-Q��_'����ᴎo��K������;��V���o��%���~OK ����*��b�f:���-ťIR��`B�5!RB@���ï�� �u �̯e\�_U�_������� g�ES��3�������QT��a����x����U<~�c?�*�#]�MW,[8O�a�x��]�1bC|踤�P��lw5V%�)�{t�<��d��5���0i�XSU��m:��Z�┵�i�"��1�^B�-��P�hJ��&)O��*�D��c�W��vM��)����}���P��ܗ-q����\mmζZ-l@�}��a��E�6��F�@��&Sg@���ݚ�M����� ȹ 4����#p�\H����dYDo�H���"��\��..R�B�H�z_�/5˘����6��KhJR��P�mƶi�m���3�,#c�co��q�a)*Pt����R�m�k�7x�D�E�\Y�閣_X�<���~�)���c[[�BP����6�Yq���S��0����%_����;��Àv�~�| VS؇ ��'O0��F0��\���U�-�d@�����7�SJ*z��3n��y��P����O���������m�~�P�3|Y��ʉr#�C�<�G~�.,! ���bqx���h~0=��!ǫ�jy����l�O,�[B��~��|9��ٱ����Xly�#�i�B��g%�S��������tˋ���e���ې��\[d�t)��.+u�|1 ������#�~Oj����hS�%��i.�~X���I�H�m��0n���c�1uE�q��cF�RF�o���7� �O�ꮧ� ���ۛ{��ʛi5�rw?׌#Qn�TW��~?y$��m\�\o����%W� ?=>S�N@�� �Ʈ���R����N�)�r"C�:��:����� �����#��qb��Y�. �6[��2K����2u�Ǧ�HYR��Q�MV��� �G�$��Q+.>�����nNH��q�^��� ����q��mM��V��D�+�-�#*�U�̒ ���p욳��u:�������IB���m���PV@O���r[b= �� ��1U�E��_Nm�yKbN�O���U�}�the�`�|6֮P>�\2�P�V���I�D�i�P�O;�9�r�mAHG�W�S]��J*�_�G��+kP�2����Ka�Z���H�'K�x�W�MZ%�O�YD�Rc+o��?�q��Ghm��d�S�oh�\�D�|:W������UA�Qc yT�q������~^�H��/��#p�CZ���T�I�1�ӏT����4��"�ČZ�����}��`w�#�*,ʹ�� ��0�i��課�Om�*�da��^gJ݅{���l�e9uF#T�ֲ��̲�ٞC"�q���ߍ ոޑ�o#�XZTp����@ o�8��(jd��xw�]�,f���`~�|,s��^����f�1���t��|��m�򸄭/ctr��5s��7�9Q�4�H1꠲BB@l9@���C�����+�wp�xu�£Yc�9��?`@#�o�mH�s2��)�=��2�.�l����jg�9$�Y�S�%*L������R�Y������7Z���,*=�䷘$�������arm�o�ϰ���UW.|�r�uf����IGw�t����Zwo��~5 ��YյhO+=8fF�)�W�7�L9lM�̘·Y���֘YLf�큹�pRF���99.A �"wz��=E\Z���'a� 2��Ǚ�#;�'}�G���*��l��^"q��+2FQ� hj��kŦ��${���ޮ-�T�٭cf�|�3#~�RJ����t��$b�(R��(����r���dx� >U b�&9,>���%E\� Ά�e�$��'�q't��*�א���ެ�b��-|d���SB�O�O��$�R+�H�)�܎�K��1m`;�J�2�Y~9��O�g8=vqD`K[�F)k�[���1m޼c��n���]s�k�z$@��)!I �x՝"v��9=�ZA=`Ɠi �:�E��)`7��vI��}d�YI�_ �o�:ob���o ���3Q��&D&�2=�� �Ά��;>�h����y.*ⅥS������Ӭ�+q&����j|UƧ����}���J0��WW< ۋS�)jQR�j���Ư��rN)�Gű�4Ѷ(�S)Ǣ�8��i��W52���No˓� ۍ%�5brOn�L�;�n��\G����=�^U�dI���8$�&���h��'���+�(������cȁ߫k�l��S^���cƗjԌE�ꭔ��gF���Ȓ��@���}O���*;e�v�WV���YJ\�]X'5��ղ�k�F��b 6R�o՜m��i N�i����>J����?��lPm�U��}>_Z&�KK��q�r��I�D�Չ~�q�3fL�:S�e>���E���-G���{L�6p�e,8��������QI��h��a�Xa��U�A'���ʂ���s�+טIjP�-��y�8ۈZ?J$��W�P� ��R�s�]��|�l(�ԓ��sƊi��o(��S0��Y� 8�T97.�����WiL��c�~�dxc�E|�2!�X�K�Ƙਫ਼�$((�6�~|d9u+�qd�^3�89��Y�6L�.I�����?���iI�q���9�)O/뚅����O���X��X�V��ZF[�یgQ�L��K1���RҖr@v�#��X�l��F���Нy�S�8�7�kF!A��sM���^rkp�jP�DyS$N���q��nxҍ!U�f�!eh�i�2�m���`�Y�I�9r�6� �TF���C}/�y�^���Η���5d�'��9A-��J��>{�_l+�`��A���[�'��յ�ϛ#w:݅�%��X�}�&�PSt�Q�"�-��\縵�/����$Ɨh�Xb�*�y��BS����;W�ջ_mc�����vt?2}1�;qS�d�d~u:2k5�2�R�~�z+|HE!)�Ǟl��7`��0�<�,�2*���Hl-��x�^����'_TV�gZA�'j� ^�2Ϊ��N7t�����?w�� �x1��f��Iz�C-Ȗ��K�^q�;���-W�DvT�7��8�Z�������� hK�(P:��Q- �8�n�Z���܃e貾�<�1�YT<�,�����"�6{/ �?�͟��|1�:�#g��W�>$����d��J��d�B��=��jf[��%rE^��il:��B���x���Sּ�1հ��,�=��*�7 fcG��#q� �eh?��2�7�����,�!7x��6�n�LC�4x��},Geǝ�tC.��vS �F�43��zz\��;QYC,6����~;RYS/6���|2���5���v��T��i����������mlv��������&� �nRh^ejR�LG�f���? �ۉҬܦƩ��|��Ȱ����>3����!v��i�ʯ�>�v��オ�X3e���_1z�Kȗ\<������!�8���V��]��?b�k41�Re��T�q��mz��TiOʦ�Z��Xq���L������q"+���2ۨ��8}�&N7XU7Ap�d�X��~�׿��&4e�o�F��� �H����O���č�c�� 懴�6���͉��+)��v;j��ݷ�� �UV�� i��� j���Y9GdÒJ1��詞�����V?h��l����l�cGs�ځ�������y�Ac�����\V3�? �� ܙg�>qH�S,�E�W�[�㺨�uch�⍸�O�}���a��>�q�6�n6����N6�q������N ! 1AQaq�0@����"2BRb�#Pr���3C`��Scst���$4D���%Td�� ?���N����a��3��m���C���w��������xA�m�q�m���m������$����4n淿t'��C"w��zU=D�\R+w�p+Y�T�&�պ@��ƃ��3ޯ?�Aﶂ��aŘ���@-�����Q�=���9D��ռ�ѻ@��M�V��P��܅�G5�f�Y<�u=,EC)�<�Fy'�"�&�չ�X~f��l�KԆV��?�� �W�N����=(� �;���{�r����ٌ�Y���h{�١������jW����P���Tc�����X�K�r��}���w�R��%��?���E��m�� �Y�q|����\lEE4���r���}�lsI�Y������f�$�=�d�yO����p�����yBj8jU�o�/�S��?�U��*������ˍ�0������u�q�m [�?f����a�� )Q�>����6#������� ?����0UQ����,IX���(6ڵ[�DI�MNލ�c&���υ�j\��X�R|,4��� j������T�hA�e��^���d���b<����n�� �즇�=!���3�^�`j�h�ȓr��jẕ�c�,ٞX����-����a�ﶔ���#�$��]w�O��Ӫ�1y%��L�Y<�wg#�ǝ�̗`�x�xa�t�w��»1���o7o5��>�m뭛C���Uƃߜ}�C���y1Xνm�F8�jI���]����H���ۺиE@I�i;r�8ӭ����V�F�Շ| ��&?�3|x�B�MuS�Ge�=Ӕ�#BE5G�����Y!z��_e��q�р/W>|-�Ci߇�t�1ޯќd�R3�u��g�=0 5��[?�#͏��q�cf���H��{ ?u�=?�?ǯ���}Z��z���hmΔ�BFTW�����<�q�(v� ��!��z���iW]*�J�V�z��gX֧A�q�&��/w���u�gYӘa���; �i=����g:��?2�dž6�ى�k�4�>�Pxs����}������G�9��3 ���)gG�R<>r h�$��'nc�h�P��Bj��J�ҧH� -��N1���N��?��~��}-q!=��_2hc�M��l�vY%UE�@|�v����M2�.Y[|y�"Eï��K�ZF,�ɯ?,q�?v�M 80jx�"�;�9vk�����+ ֧�� �ȺU��?�%�vcV��mA�6��Qg^M����A}�3�nl� QRN�l8�kkn�'�����(��M�7m9و�q���%ޟ���*h$Zk"��$�9��: �?U8�Sl��,,|ɒ��xH(ѷ����Gn�/Q�4�P��G�%��Ա8�N��!� �&�7�;���eKM7�4��9R/%����l�c>�x;������>��C�:�����t��h?aKX�bhe�ᜋ^�$�Iհ �hr7%F$�E��Fd���t��5���+�(M6�t����Ü�UU|zW�=a�Ts�Tg������dqP�Q����b'�m���1{|Y����X�N��b �P~��F^F:����k6�"�j!�� �I�r�`��1&�-$�Bevk:y���#yw��I0��x��=D�4��tU���P�ZH��ڠ底taP��6����b>�xa����Q�#� WeF��ŮNj�p�J* mQ�N����*I�-*�ȩ�F�g�3 �5��V�ʊ�ɮ�a��5F���O@{���NX��?����H�]3��1�Ri_u��������ѕ�� ����0��� F��~��:60�p�͈�S��qX#a�5>���`�o&+�<2�D����: �������ڝ�$�nP���*)�N�|y�Ej�F�5ټ�e���ihy�Z �>���k�bH�a�v��h�-#���!�Po=@k̆IEN��@��}Ll?j�O������߭�ʞ���Q|A07x���wt!xf���I2?Z��<ץ�T���cU�j��]��陎Ltl �}5�ϓ��$�,��O�mˊ�;�@O��jE��j(�ا,��LX���LO���Ц�90�O �.����a��nA���7������j4 ��W��_ٓ���zW�jcB������y՗+EM�)d���N�g6�y1_x��p�$Lv:��9�"z��p���ʙ$��^��JԼ*�ϭ����o���=x�Lj�6�J��u82�A�H�3$�ٕ@�=Vv�]�'�qEz�;I˼��)��=��ɯ���x �/�W(V���p�����$ �m�������u�����񶤑Oqˎ�T����r��㠚x�sr�GC��byp�G��1ߠ�w e�8�$⿄����/�M{*}��W�]˷.�CK\�ުx���/$�WPw���r� |i���&�}�{�X� �>��$-��l���?-z���g����lΆ���(F���h�vS*���b���߲ڡn,|)mrH[���a�3�ר�[1��3o_�U�3�TC�$��(�=�)0�kgP���� ��u�^=��4 �WYCҸ:��vQ�ר�X�à��tk�m,�t*��^�,�}D*� �"(�I��9R����>`�`��[~Q]�#af��i6l��8���6�:,s�s�N6�j"�A4���IuQ��6E,�GnH��zS�HO�uk�5$�I�4��ؤ�Q9�@��C����wp�BGv[]�u�Ov���0I4���\��y�����Q�Ѹ��~>Z��8�T��a��q�ޣ;z��a���/��S��I:�ܫ_�|������>=Z����8:�S��U�I�J��"IY���8%b8���H��:�QO�6�;7�I�S��J��ҌAά3��>c���E+&jf$eC+�z�;��V����� �r���ʺ������my�e���aQ�f&��6�ND��.:��NT�vm�<- u���ǝ\MvZY�N�NT��-A�>jr!S��n�O 1�3�Ns�%�3D@���`������ܟ 1�^c<���� �a�ɽ�̲�Xë#�w�|y�cW�=�9I*H8�p�^(4���՗�k��arOcW�tO�\�ƍR��8����'�K���I�Q�����?5�>[�}��yU�ײ -h��=��% q�ThG�2�)���"ו3]�!kB��*p�FDl�A���,�eEi�H�f�Ps�����5�H:�Փ~�H�0Dت�D�I����h�F3�������c��2���E��9�H��5�zԑ�ʚ�i�X�=:m�xg�hd(�v����׊�9iS��O��d@0ڽ���:�p�5�h-��t�&���X�q�ӕ,��ie�|���7A�2���O%P��E��htj��Y1��w�Ѓ!����  ���� ࢽ��My�7�\�a�@�ţ�J �4�Ȼ�F�@o�̒?4�wx��)��]�P��~�����u�����5�����7X ��9��^ܩ�U;Iꭆ 5 �������eK2�7(�{|��Y׎ �V��\"���Z�1� Z�����}��(�Ǝ"�1S���_�vE30>���p;� ΝD��%x�W�?W?v����o�^V�i�d��r[��/&>�~`�9Wh��y�;���R��� ;;ɮT��?����r$�g1�K����A��C��c��K��l:�'��3 c�ﳯ*"t8�~l��)���m��+U,z��`(�>yJ�?����h>��]��v��ЍG*�{`��;y]��I�T� ;c��NU�fo¾h���/$���|NS���1�S�"�H��V���T���4��uhǜ�]�v;���5�͠x��'C\�SBpl���h}�N����� A�Bx���%��ޭ�l��/����T��w�ʽ]D�=����K���ž�r㻠l4�S�O?=�k �M:� ��c�C�a�#ha���)�ѐxc�s���gP�iG��{+���x���Q���I= �� z��ԫ+ �8"�k�ñ�j=|����c ��y��CF��/��*9ж�h{ �?4�o� ��k�m�Q�N�x��;�Y��4膚�a�w?�6�>e]�����Q�r�:����g�,i"�����ԩA�*M�<�G��b�if��l^M��5� �Ҩ�{����6J��ZJ�����P�*�����Y���ݛu�_4�9�I8�7���������,^ToR���m4�H��?�N�S�ѕw��/S��甍�@�9H�S�T��t�ƻ���ʒU��*{Xs�@����f�����֒Li�K{H�w^���������Ϥm�tq���s� ���ք��f:��o~s��g�r��ט� �S�ѱC�e]�x���a��) ���(b-$(�j>�7q�B?ӕ�F��hV25r[7 Y� }L�R��}����*sg+��x�r�2�U=�*'WS��ZDW]�WǞ�<��叓���{�$�9Ou4��y�90-�1�'*D`�c�^o?(�9��u���ݐ��'PI&� f�Jݮ�������:wS����jfP1F:X �H�9dԯ���˝[�_54 �}*;@�ܨ�� ð�yn�T���?�ןd�#���4rG�ͨ��H�1�|-#���Mr�S3��G�3�����)�.᧏3v�z֑��r����$G"�`j �1t��x0<Ɔ�Wh6�y�6��,œ�Ga��gA����y��b��)��h�D��ß�_�m��ü �gG;��e�v��ݝ�nQ� ��C����-�*��o���y�a��M��I�>�<���]obD��"�:���G�A��-\%LT�8���c�)��+y76���o�Q�#*{�(F�⽕�y����=���rW�\p���۩�c���A���^e6��K������ʐ�cVf5$�'->���ՉN"���F�"�UQ@�f��Gb~��#�&�M=��8�ט�JNu9��D��[̤�s�o�~������ G��9T�tW^g5y$b��Y'��س�Ǵ�=��U-2 #�MC�t(�i� �lj�@Q 5�̣i�*�O����s�x�K�f��}\��M{E�V�{�υ��Ƈ�����);�H����I��fe�Lȣr�2��>��W�I�Ȃ6������i��k�� �5�YOxȺ����>��Y�f5'��|��H+��98pj�n�.O�y�������jY��~��i�w'������l�;�s�2��Y��:'lg�ꥴ)o#'Sa�a�K��Z� �m��}�`169�n���"���x��I ��*+� }F<��cГ���F�P�������ֹ*�PqX�x۩��,� ��N�� �4<-����%����:��7����W���u�`����� $�?�I��&����o��o��`v�>��P��"��l���4��5'�Z�gE���8���?��[�X�7(��.Q�-��*���ތL@̲����v��.5���[��=�t\+�CNܛ��,g�SQnH����}*F�G16���&:�t��4ُ"A��̣��$�b �|����#rs��a�����T�� ]�<�j��BS�('$�ɻ� �wP;�/�n��?�ݜ��x�F��yUn�~mL*-�������Xf�wd^�a�}��f�,=t�׵i�.2/wpN�Ep8�OР���•��R�FJ� 55TZ��T �ɭ�<��]��/�0�r�@�f��V��V����Nz�G��^���7hZi����k��3�,kN�e|�vg�1{9]_i��X5y7� 8e]�U����'�-2,���e"����]ot�I��Y_��n�(JҼ��1�O ]bXc���Nu�No��pS���Q_���_�?i�~�x h5d'�(qw52] ��'ޤ�q��o1�R!���`ywy�A4u���h<קy���\[~�4�\ X�Wt/� 6�����n�F�a8��f���z �3$�t(���q��q�x��^�XWeN'p<-v�!�{�(>ӽDP7��ո0�y)�e$ٕv�Ih'Q�EA�m*�H��RI��=:��� ���4牢) �%_iN�ݧ�l]� �Nt���G��H�L��� ɱ�g<���1V�,�J~�ٹ�"K��Q�� 9�HS�9�?@��k����r�;we݁�]I�!{ �@�G�[�"��`���J:�n]�{�cA�E����V��ʆ���#��U9�6����j�#Y�m\��q�e4h�B�7��C�������d<�?J����1g:ٳ���=Y���D�p�ц� ׈ǔ��1�]26؜oS�'��9�V�FVu�P�h�9�xc�oq�X��p�o�5��Ա5$�9W�V(�[Ak�aY錎qf;�'�[�|���b�6�Ck��)��#a#a˙��8���=äh�4��2��C��4tm^ �n'c���]GQ$[Wҿ��i���vN�{Fu ��1�gx��1┷���N�m��{j-,��x�� Ūm�ЧS�[�s���Gna���䑴�� x�p 8<������97�Q���ϴ�v�aϚG��Rt�Һ׈�f^\r��WH�JU�7Z���y)�vg=����n��4�_)y��D'y�6�]�c�5̪�\� �PF�k����&�c;��cq�$~T�7j ���nç]�<�g ":�to�t}�159�<�/�8������m�b�K#g'I'.W�����6��I/��>v��\�MN��g���m�A�yQL�4u�Lj�j9��#44�t��l^�}L����n��R��!��t��±]��r��h6ٍ>�yҏ�N��fU�� ���� Fm@�8}�/u��jb9������he:A�y�ծw��GpΧh�5����l}�3p468��)U��d��c����;Us/�֔�YX�1�O2��uq�s��`hwg�r~�{ R��mhN��؎*q 42�*th��>�#���E����#��Hv�O����q�}�����6�e��\�,Wk�#���X��b>��p}�դ��3���T5��†��6��[��@�P�y*n��|'f�֧>�lư΂�̺����SU�'*�q�p�_S�����M�� '��c�6�����m�� ySʨ;M��r���Ƌ�m�Kxo,���Gm�P��A�G�:��i��w�9�}M(�^�V��$ǒ�ѽ�9���|���� �a����J�SQ�a���r�B;����}���ٻ֢�2�%U���c�#�g���N�a�ݕ�'�v�[�OY'��3L�3�;,p�]@�S��{ls��X�'���c�jw�k'a�.��}�}&�� �dP�*�bK=ɍ!����;3n�gΊU�ߴmt�'*{,=SzfD� A��ko~�G�aoq�_mi}#�m�������P�Xhύ����mxǍ�΂���巿zf��Q���c���|kc�����?���W��Y�$���_Lv����l߶��c���`?����l�j�ݲˏ!V��6����U�Ђ(A���4y)H���p�Z_�x��>���e��R��$�/�`^'3qˏ�-&Q�=?��CFVR �D�fV�9��{�8g�������n�h�(P"��6�[�D���< E�����~0<@�`�G�6����Hг�cc�� �c�K.5��D��d�B���`?�XQ��2��ٿyqo&+�1^� DW�0�ꊩ���G�#��Q�nL3��c���������/��x ��1�1[y�x�პCW��C�c�UĨ80�m�e�4.{�m��u���I=��f�����0QRls9���f���������9���~f�����Ǩ��a�"@�8���ȁ�Q����#c�ic������G��$���G���r/$W�(��W���V�"��m�7�[m�A�m����bo��D� j����۳� l���^�k�h׽����� ��#� iXn�v��eT�k�a�^Y�4�BN��ĕ��0 !01@Q"2AaPq3BR������?���@4�Q�����T3,���㺠�W�[=JK�Ϟ���2�r^7��vc�:�9 �E�ߴ�w�S#d���Ix��u��:��Hp��9E!�� V 2;73|F��9Y���*ʬ�F��D����u&���y؟��^EA��A��(ɩ���^��GV:ݜDy�`��Jr29ܾ�㝉��[���E;Fzx��YG��U�e�Y�C���� ����v-tx����I�sם�Ę�q��Eb�+P\ :>�i�C'�;�����k|z�رn�y]�#ǿb��Q��������w�����(�r|ӹs��[�D��2v-%��@;�8<a���[\o[ϧw��I!��*0�krs)�[�J9^��ʜ��p1)� "��/_>��o��<1����A�E�y^�C��`�x1'ܣn�p��s`l���fQ��):�l����b>�Me�jH^?�kl3(�z:���1ŠK&?Q�~�{�ٺ�h�y���/�[��V�|6��}�KbX����mn[-��7�5q�94�������dm���c^���h� X��5��<�eޘ>G���-�}�دB�ޟ� ��|�rt�M��V+�]�c?�-#ڛ��^ǂ}���Lkr���O��u�>�-D�ry� D?:ޞ�U��ǜ�7�V��?瓮�"�#���r��չģVR;�n���/_� ؉v�ݶe5d�b9��/O��009�G���5n�W����JpA�*�r9�>�1��.[t���s�F���nQ� V 77R�]�ɫ8����_0<՜�IF�u(v��4��F�k�3��E)��N:��yڮe��P�`�1}�$WS��J�SQ�N�j�ٺ��޵�#l���ј(�5=��5�lǏmoW�v-�1����v,W�mn��߀$x�<����v�j(����c]��@#��1������Ǔ���o'��u+����;G�#�޸��v-lη��/(`i⣍Pm^���ԯ̾9Z��F��������n��1��� ��]�[��)�'������:�֪�W��FC����� �B9،!?���]��V��A�Վ�M��b�w��G F>_DȬ0¤�#�QR�[V��kz���m�w�"��9ZG�7'[��=�Q����j8R?�zf�\a�=��O�U����*oB�A�|G���2�54 �p��.w7� �� ��&������ξxGHp� B%��$g�����t�Џ򤵍z���HN�u�Я�-�'4��0��;_��3 !01"@AQa2Pq#3BR������?��ʩca��en��^��8���<�u#��m*08r��y�N"�<�Ѳ0��@\�p��� �����Kv�D��J8�Fҽ� �f�Y��-m�ybX�NP����}�!*8t(�OqѢ��Q�wW�K��ZD��Δ^e��!� ��B�K��p~�����e*l}z#9ң�k���q#�Ft�o��S�R����-�w�!�S���Ӥß|M�l޶V��!eˈ�8Y���c�ЮM2��tk���� ������J�fS����Ö*i/2�����n]�k�\���|4yX�8��U�P.���Ы[���l��@"�t�<������5�lF���vU�����W��W��;�b�cД^6[#7@vU�xgZv��F�6��Q,K�v��� �+Ъ��n��Ǣ��Ft���8��0��c�@�!�Zq s�v�t�;#](B��-�nῃ~���3g������5�J�%���O������n�kB�ĺ�.r��+���#�N$?�q�/�s�6��p��a����a��J/��M�8��6�ܰ"�*������ɗud"\w���aT(����[��F��U՛����RT�b���n�*��6���O��SJ�.�ij<�v�MT��R\c��5l�sZB>F��<7�;EA��{��E���Ö��1U/�#��d1�a�n.1ě����0�ʾR�h��|�R��Ao�3�m3 ��%�� ���28Q� ��y��φ���H�To�7�lW>����#i`�q���c����a��� �m,B�-j����݋�'mR1Ήt�>��V��p���s�0IbI�C.���1R�ea�����]H�6����������4B>��o��](��$B���m�����a�!=��?�B� K�Ǿ+�Ծ"�n���K��*��+��[T#�{E�J�S����Q�����s�5�:�U�\wĐ�f�3����܆&�)����I���Ԇw��E T�lrTf6Q|R�h:��[K�� �z��c֧�G�C��%\��_�a�84��HcO�bi��ؖV��7H �)*ģK~Xhչ0��4?�0��� �E<���}3���#���u�?�� ��|g�S�6ꊤ�|�I#Hڛ� �ա��w�X��9��7���Ŀ%�SL��y6č��|�F�a 8���b��$�sק�h���b9RAu7�˨p�Č�_\*w��묦��F ����4D~�f����|(�"m���NK��i�S�>�$d7SlA��/�²����SL��|6N�}���S�˯���g��]6��; �#�.��<���q'Q�1|KQ$�����񛩶"�$r�b:���N8�w@��8$�� �AjfG|~�9F ���Y��ʺ��Bwؒ������M:I岎�G��`s�YV5����6��A �b:�W���G�q%l�����F��H���7�������Fsv7��k�� 403WebShell
403Webshell
Server IP : 213.165.242.4  /  Your IP : 216.73.216.54
Web Server : Apache
System : Linux amsngx344.inmotionhosting.com 4.18.0-553.40.1.lve.el8.x86_64 #1 SMP Wed Feb 12 18:54:57 UTC 2025 x86_64
User : aquafi9 ( 1305)
PHP Version : 7.4.33
Disable Function : NONE
MySQL : OFF  |  cURL : ON  |  WGET : ON  |  Perl : ON  |  Python : ON  |  Sudo : ON  |  Pkexec : ON
Directory :  /lib/clang/20/include/

Upload File :
current_dir [ Writeable ] document_root [ Writeable ]

 

Command :

[ Back ]     

Current File : /lib/clang/20/include/xmmintrin.h
/*===---- xmmintrin.h - SSE intrinsics -------------------------------------===
 *
 * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 * See https://llvm.org/LICENSE.txt for license information.
 * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 *
 *===-----------------------------------------------------------------------===
 */

#ifndef __XMMINTRIN_H
#define __XMMINTRIN_H

#if !defined(__i386__) && !defined(__x86_64__)
#error "This header is only meant to be used on x86 and x64 architecture"
#endif

#include <mmintrin.h>

typedef int __v4si __attribute__((__vector_size__(16)));
typedef float __v4sf __attribute__((__vector_size__(16)));
typedef float __m128 __attribute__((__vector_size__(16), __aligned__(16)));

typedef float __m128_u __attribute__((__vector_size__(16), __aligned__(1)));

/* Unsigned types */
typedef unsigned int __v4su __attribute__((__vector_size__(16)));

/* This header should only be included in a hosted environment as it depends on
 * a standard library to provide allocation routines. */
#if __STDC_HOSTED__
#include <mm_malloc.h>
#endif

/* Define the default attributes for the functions in this file. */
#if defined(__EVEX512__) && !defined(__AVX10_1_512__)
#define __DEFAULT_FN_ATTRS                                                     \
  __attribute__((__always_inline__, __nodebug__, __target__("sse,no-evex512"), \
                 __min_vector_width__(128)))
#define __DEFAULT_FN_ATTRS_SSE2                                                \
  __attribute__((__always_inline__, __nodebug__,                               \
                 __target__("sse2,no-evex512"), __min_vector_width__(128)))
#else
#define __DEFAULT_FN_ATTRS                                                     \
  __attribute__((__always_inline__, __nodebug__, __target__("sse"),            \
                 __min_vector_width__(128)))
#define __DEFAULT_FN_ATTRS_SSE2                                                \
  __attribute__((__always_inline__, __nodebug__, __target__("sse2"),           \
                 __min_vector_width__(128)))
#endif

#if defined(__cplusplus) && (__cplusplus >= 201103L)
#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS constexpr
#define __DEFAULT_FN_ATTRS_SSE2_CONSTEXPR __DEFAULT_FN_ATTRS_SSE2 constexpr
#else
#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS
#define __DEFAULT_FN_ATTRS_SSE2_CONSTEXPR __DEFAULT_FN_ATTRS_SSE2
#endif

#define __trunc64(x)                                                           \
  (__m64) __builtin_shufflevector((__v2di)(x), __extension__(__v2di){}, 0)
#define __zext128(x)                                                           \
  (__m128i) __builtin_shufflevector((__v2si)(x), __extension__(__v2si){}, 0,   \
                                    1, 2, 3)
#define __anyext128(x)                                                         \
  (__m128i) __builtin_shufflevector((__v2si)(x), __extension__(__v2si){}, 0,   \
                                    1, -1, -1)
#define __zeroupper64(x)                                                       \
  (__m128i) __builtin_shufflevector((__v4si)(x), __extension__(__v4si){}, 0,   \
                                    1, 4, 5)

/// Adds the 32-bit float values in the low-order bits of the operands.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VADDSS / ADDSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the source operands.
///    The lower 32 bits of this operand are used in the calculation.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the source operands.
///    The lower 32 bits of this operand are used in the calculation.
/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the sum
///    of the lower 32 bits of both operands. The upper 96 bits are copied from
///    the upper 96 bits of the first source operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_add_ss(__m128 __a, __m128 __b) {
  __a[0] += __b[0];
  return __a;
}

/// Adds two 128-bit vectors of [4 x float], and returns the results of
///    the addition.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VADDPS / ADDPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the source operands.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the source operands.
/// \returns A 128-bit vector of [4 x float] containing the sums of both
///    operands.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_add_ps(__m128 __a, __m128 __b) {
  return (__m128)((__v4sf)__a + (__v4sf)__b);
}

/// Subtracts the 32-bit float value in the low-order bits of the second
///    operand from the corresponding value in the first operand.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VSUBSS / SUBSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing the minuend. The lower 32 bits
///    of this operand are used in the calculation.
/// \param __b
///    A 128-bit vector of [4 x float] containing the subtrahend. The lower 32
///    bits of this operand are used in the calculation.
/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
///    difference of the lower 32 bits of both operands. The upper 96 bits are
///    copied from the upper 96 bits of the first source operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_sub_ss(__m128 __a, __m128 __b) {
  __a[0] -= __b[0];
  return __a;
}

/// Subtracts each of the values of the second operand from the first
///    operand, both of which are 128-bit vectors of [4 x float] and returns
///    the results of the subtraction.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VSUBPS / SUBPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing the minuend.
/// \param __b
///    A 128-bit vector of [4 x float] containing the subtrahend.
/// \returns A 128-bit vector of [4 x float] containing the differences between
///    both operands.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_sub_ps(__m128 __a, __m128 __b) {
  return (__m128)((__v4sf)__a - (__v4sf)__b);
}

/// Multiplies two 32-bit float values in the low-order bits of the
///    operands.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMULSS / MULSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the source operands.
///    The lower 32 bits of this operand are used in the calculation.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the source operands.
///    The lower 32 bits of this operand are used in the calculation.
/// \returns A 128-bit vector of [4 x float] containing the product of the lower
///    32 bits of both operands. The upper 96 bits are copied from the upper 96
///    bits of the first source operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_mul_ss(__m128 __a, __m128 __b) {
  __a[0] *= __b[0];
  return __a;
}

/// Multiplies two 128-bit vectors of [4 x float] and returns the
///    results of the multiplication.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMULPS / MULPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the source operands.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the source operands.
/// \returns A 128-bit vector of [4 x float] containing the products of both
///    operands.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_mul_ps(__m128 __a, __m128 __b) {
  return (__m128)((__v4sf)__a * (__v4sf)__b);
}

/// Divides the value in the low-order 32 bits of the first operand by
///    the corresponding value in the second operand.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VDIVSS / DIVSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing the dividend. The lower 32
///    bits of this operand are used in the calculation.
/// \param __b
///    A 128-bit vector of [4 x float] containing the divisor. The lower 32 bits
///    of this operand are used in the calculation.
/// \returns A 128-bit vector of [4 x float] containing the quotients of the
///    lower 32 bits of both operands. The upper 96 bits are copied from the
///    upper 96 bits of the first source operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_div_ss(__m128 __a, __m128 __b) {
  __a[0] /= __b[0];
  return __a;
}

/// Divides two 128-bit vectors of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VDIVPS / DIVPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing the dividend.
/// \param __b
///    A 128-bit vector of [4 x float] containing the divisor.
/// \returns A 128-bit vector of [4 x float] containing the quotients of both
///    operands.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_div_ps(__m128 __a, __m128 __b) {
  return (__m128)((__v4sf)__a / (__v4sf)__b);
}

/// Calculates the square root of the value stored in the low-order bits
///    of a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VSQRTSS / SQRTSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the calculation.
/// \returns A 128-bit vector of [4 x float] containing the square root of the
///    value in the low-order bits of the operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_sqrt_ss(__m128 __a)
{
  return (__m128)__builtin_ia32_sqrtss((__v4sf)__a);
}

/// Calculates the square roots of the values stored in a 128-bit vector
///    of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VSQRTPS / SQRTPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the square roots of the
///    values in the operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_sqrt_ps(__m128 __a)
{
  return __builtin_ia32_sqrtps((__v4sf)__a);
}

/// Calculates the approximate reciprocal of the value stored in the
///    low-order bits of a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VRCPSS / RCPSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the calculation.
/// \returns A 128-bit vector of [4 x float] containing the approximate
///    reciprocal of the value in the low-order bits of the operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_rcp_ss(__m128 __a)
{
  return (__m128)__builtin_ia32_rcpss((__v4sf)__a);
}

/// Calculates the approximate reciprocals of the values stored in a
///    128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VRCPPS / RCPPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the approximate
///    reciprocals of the values in the operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_rcp_ps(__m128 __a)
{
  return (__m128)__builtin_ia32_rcpps((__v4sf)__a);
}

/// Calculates the approximate reciprocal of the square root of the value
///    stored in the low-order bits of a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VRSQRTSS / RSQRTSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the calculation.
/// \returns A 128-bit vector of [4 x float] containing the approximate
///    reciprocal of the square root of the value in the low-order bits of the
///    operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_rsqrt_ss(__m128 __a)
{
  return __builtin_ia32_rsqrtss((__v4sf)__a);
}

/// Calculates the approximate reciprocals of the square roots of the
///    values stored in a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VRSQRTPS / RSQRTPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the approximate
///    reciprocals of the square roots of the values in the operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_rsqrt_ps(__m128 __a)
{
  return __builtin_ia32_rsqrtps((__v4sf)__a);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands and returns the lesser value in the low-order bits of the
///    vector of [4 x float].
///
///    If either value in a comparison is NaN, returns the value from \a __b.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMINSS / MINSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
///    minimum value between both operands. The upper 96 bits are copied from
///    the upper 96 bits of the first source operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_min_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_minss((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 128-bit vectors of [4 x float] and returns the lesser
///    of each pair of values.
///
///    If either value in a comparison is NaN, returns the value from \a __b.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMINPS / MINPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands.
/// \returns A 128-bit vector of [4 x float] containing the minimum values
///    between both operands.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_min_ps(__m128 __a, __m128 __b)
{
  return __builtin_ia32_minps((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands and returns the greater value in the low-order bits of a 128-bit
///    vector of [4 x float].
///
///    If either value in a comparison is NaN, returns the value from \a __b.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMAXSS / MAXSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
///    maximum value between both operands. The upper 96 bits are copied from
///    the upper 96 bits of the first source operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_max_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_maxss((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 128-bit vectors of [4 x float] and returns the greater
///    of each pair of values.
///
///    If either value in a comparison is NaN, returns the value from \a __b.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMAXPS / MAXPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands.
/// \returns A 128-bit vector of [4 x float] containing the maximum values
///    between both operands.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_max_ps(__m128 __a, __m128 __b)
{
  return __builtin_ia32_maxps((__v4sf)__a, (__v4sf)__b);
}

/// Performs a bitwise AND of two 128-bit vectors of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VANDPS / ANDPS </c> instructions.
///
/// \param __a
///    A 128-bit vector containing one of the source operands.
/// \param __b
///    A 128-bit vector containing one of the source operands.
/// \returns A 128-bit vector of [4 x float] containing the bitwise AND of the
///    values between both operands.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_and_ps(__m128 __a, __m128 __b) {
  return (__m128)((__v4su)__a & (__v4su)__b);
}

/// Performs a bitwise AND of two 128-bit vectors of [4 x float], using
///    the one's complement of the values contained in the first source
///    operand.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VANDNPS / ANDNPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing the first source operand. The
///    one's complement of this value is used in the bitwise AND.
/// \param __b
///    A 128-bit vector of [4 x float] containing the second source operand.
/// \returns A 128-bit vector of [4 x float] containing the bitwise AND of the
///    one's complement of the first operand and the values in the second
///    operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_andnot_ps(__m128 __a, __m128 __b) {
  return (__m128)(~(__v4su)__a & (__v4su)__b);
}

/// Performs a bitwise OR of two 128-bit vectors of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VORPS / ORPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the source operands.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the source operands.
/// \returns A 128-bit vector of [4 x float] containing the bitwise OR of the
///    values between both operands.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_or_ps(__m128 __a, __m128 __b) {
  return (__m128)((__v4su)__a | (__v4su)__b);
}

/// Performs a bitwise exclusive OR of two 128-bit vectors of
///    [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VXORPS / XORPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the source operands.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the source operands.
/// \returns A 128-bit vector of [4 x float] containing the bitwise exclusive OR
///    of the values between both operands.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_xor_ps(__m128 __a, __m128 __b) {
  return (__m128)((__v4su)__a ^ (__v4su)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands for equality.
///
///    The comparison returns 0x0 for false, 0xFFFFFFFF for true, in the
///    low-order bits of a vector [4 x float].
///    If either value in a comparison is NaN, returns false.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPEQSS / CMPEQSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpeq_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpeqss((__v4sf)__a, (__v4sf)__b);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] for equality.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true.
///    If either value in a comparison is NaN, returns false.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPEQPS / CMPEQPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpeq_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpeqps((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the value in the first operand is less than the
///    corresponding value in the second operand.
///
///    The comparison returns 0x0 for false, 0xFFFFFFFF for true, in the
///    low-order bits of a vector of [4 x float].
///    If either value in a comparison is NaN, returns false.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPLTSS / CMPLTSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmplt_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpltss((__v4sf)__a, (__v4sf)__b);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] to determine if the values in the first
///    operand are less than those in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
///    If either value in a comparison is NaN, returns false.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPLTPS / CMPLTPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmplt_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpltps((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the value in the first operand is less than or
///    equal to the corresponding value in the second operand.
///
///    The comparison returns 0x0 for false, 0xFFFFFFFFFFFFFFFF for true, in
///    the low-order bits of a vector of [4 x float].
///    If either value in a comparison is NaN, returns false.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPLESS / CMPLESS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmple_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpless((__v4sf)__a, (__v4sf)__b);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] to determine if the values in the first
///    operand are less than or equal to those in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true.
///    If either value in a comparison is NaN, returns false.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPLEPS / CMPLEPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmple_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpleps((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the value in the first operand is greater than
///    the corresponding value in the second operand.
///
///    The comparison returns 0x0 for false, 0xFFFFFFFF for true, in the
///    low-order bits of a vector of [4 x float].
///    If either value in a comparison is NaN, returns false.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPLTSS / CMPLTSS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpgt_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_shufflevector((__v4sf)__a,
                                         (__v4sf)__builtin_ia32_cmpltss((__v4sf)__b, (__v4sf)__a),
                                         4, 1, 2, 3);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] to determine if the values in the first
///    operand are greater than those in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true.
///    If either value in a comparison is NaN, returns false.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPLTPS / CMPLTPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpgt_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpltps((__v4sf)__b, (__v4sf)__a);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the value in the first operand is greater than
///    or equal to the corresponding value in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true, in the
///    low-order bits of a vector of [4 x float].
///    If either value in a comparison is NaN, returns false.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPLESS / CMPLESS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpge_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_shufflevector((__v4sf)__a,
                                         (__v4sf)__builtin_ia32_cmpless((__v4sf)__b, (__v4sf)__a),
                                         4, 1, 2, 3);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] to determine if the values in the first
///    operand are greater than or equal to those in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
///    If either value in a comparison is NaN, returns false.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPLEPS / CMPLEPS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpge_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpleps((__v4sf)__b, (__v4sf)__a);
}

/// Compares two 32-bit float values in the low-order bits of both operands
///    for inequality.
///
///    The comparison returns 0x0 for false, 0xFFFFFFFF for true, in the
///    low-order bits of a vector of [4 x float].
///    If either value in a comparison is NaN, returns true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPNEQSS / CMPNEQSS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpneq_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpneqss((__v4sf)__a, (__v4sf)__b);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] for inequality.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true.
///    If either value in a comparison is NaN, returns true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPNEQPS / CMPNEQPS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpneq_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpneqps((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the value in the first operand is not less than
///    the corresponding value in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true, in the
///    low-order bits of a vector of [4 x float].
///    If either value in a comparison is NaN, returns true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPNLTSS / CMPNLTSS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpnlt_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpnltss((__v4sf)__a, (__v4sf)__b);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] to determine if the values in the first
///    operand are not less than those in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true.
///    If either value in a comparison is NaN, returns true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPNLTPS / CMPNLTPS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpnlt_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpnltps((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the value in the first operand is not less than
///    or equal to the corresponding value in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true, in the
///    low-order bits of a vector of [4 x float].
///    If either value in a comparison is NaN, returns true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPNLESS / CMPNLESS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpnle_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpnless((__v4sf)__a, (__v4sf)__b);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] to determine if the values in the first
///    operand are not less than or equal to those in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true.
///    If either value in a comparison is NaN, returns true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPNLEPS / CMPNLEPS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpnle_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpnleps((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the value in the first operand is not greater
///    than the corresponding value in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true, in the
///    low-order bits of a vector of [4 x float].
///    If either value in a comparison is NaN, returns true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPNLTSS / CMPNLTSS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpngt_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_shufflevector((__v4sf)__a,
                                         (__v4sf)__builtin_ia32_cmpnltss((__v4sf)__b, (__v4sf)__a),
                                         4, 1, 2, 3);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] to determine if the values in the first
///    operand are not greater than those in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true.
///    If either value in a comparison is NaN, returns true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPNLTPS / CMPNLTPS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpngt_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpnltps((__v4sf)__b, (__v4sf)__a);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the value in the first operand is not greater
///    than or equal to the corresponding value in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true, in the
///    low-order bits of a vector of [4 x float].
///    If either value in a comparison is NaN, returns true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPNLESS / CMPNLESS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpnge_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_shufflevector((__v4sf)__a,
                                         (__v4sf)__builtin_ia32_cmpnless((__v4sf)__b, (__v4sf)__a),
                                         4, 1, 2, 3);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] to determine if the values in the first
///    operand are not greater than or equal to those in the second operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true.
///    If either value in a comparison is NaN, returns true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPNLEPS / CMPNLEPS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpnge_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpnleps((__v4sf)__b, (__v4sf)__a);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the value in the first operand is ordered with
///    respect to the corresponding value in the second operand.
///
///    A pair of floating-point values are ordered with respect to each
///    other if neither value is a NaN. Each comparison returns 0x0 for false,
///    0xFFFFFFFF for true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPORDSS / CMPORDSS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpord_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpordss((__v4sf)__a, (__v4sf)__b);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] to determine if the values in the first
///    operand are ordered with respect to those in the second operand.
///
///    A pair of floating-point values are ordered with respect to each
///    other if neither value is a NaN. Each comparison returns 0x0 for false,
///    0xFFFFFFFF for true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPORDPS / CMPORDPS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpord_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpordps((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the value in the first operand is unordered
///    with respect to the corresponding value in the second operand.
///
///    A pair of double-precision values are unordered with respect to each
///    other if one or both values are NaN. Each comparison returns 0x0 for
///    false, 0xFFFFFFFF for true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPUNORDSS / CMPUNORDSS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float] containing one of the operands. The lower
///    32 bits of this operand are used in the comparison.
/// \returns A 128-bit vector of [4 x float] containing the comparison results
///    in the low-order bits.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpunord_ss(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpunordss((__v4sf)__a, (__v4sf)__b);
}

/// Compares each of the corresponding 32-bit float values of the
///    128-bit vectors of [4 x float] to determine if the values in the first
///    operand are unordered with respect to those in the second operand.
///
///    A pair of double-precision values are unordered with respect to each
///    other if one or both values are NaN. Each comparison returns 0x0 for
///    false, 0xFFFFFFFFFFFFFFFF for true.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCMPUNORDPS / CMPUNORDPS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 128-bit vector of [4 x float].
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_cmpunord_ps(__m128 __a, __m128 __b)
{
  return (__m128)__builtin_ia32_cmpunordps((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands for equality.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_comieq_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_comieq((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the first operand is less than the second
///    operand.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_comilt_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_comilt((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the first operand is less than or equal to the
///    second operand.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_comile_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_comile((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the first operand is greater than the second
///    operand.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_comigt_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_comigt((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the first operand is greater than or equal to
///    the second operand.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_comige_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_comige((__v4sf)__a, (__v4sf)__b);
}

/// Compares two 32-bit float values in the low-order bits of both
///    operands to determine if the first operand is not equal to the second
///    operand.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 1.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_comineq_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_comineq((__v4sf)__a, (__v4sf)__b);
}

/// Performs an unordered comparison of two 32-bit float values using
///    the low-order bits of both operands to determine equality.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_ucomieq_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_ucomieq((__v4sf)__a, (__v4sf)__b);
}

/// Performs an unordered comparison of two 32-bit float values using
///    the low-order bits of both operands to determine if the first operand is
///    less than the second operand.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_ucomilt_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_ucomilt((__v4sf)__a, (__v4sf)__b);
}

/// Performs an unordered comparison of two 32-bit float values using
///    the low-order bits of both operands to determine if the first operand is
///    less than or equal to the second operand.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_ucomile_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_ucomile((__v4sf)__a, (__v4sf)__b);
}

/// Performs an unordered comparison of two 32-bit float values using
///    the low-order bits of both operands to determine if the first operand is
///    greater than the second operand.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_ucomigt_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_ucomigt((__v4sf)__a, (__v4sf)__b);
}

/// Performs an unordered comparison of two 32-bit float values using
///    the low-order bits of both operands to determine if the first operand is
///    greater than or equal to the second operand.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_ucomige_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_ucomige((__v4sf)__a, (__v4sf)__b);
}

/// Performs an unordered comparison of two 32-bit float values using
///    the low-order bits of both operands to determine inequality.
///
///    The comparison returns 0 for false, 1 for true. If either value in a
///    comparison is NaN, returns 0.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \param __b
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the comparison.
/// \returns An integer containing the comparison results.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_ucomineq_ss(__m128 __a, __m128 __b)
{
  return __builtin_ia32_ucomineq((__v4sf)__a, (__v4sf)__b);
}

/// Converts a float value contained in the lower 32 bits of a vector of
///    [4 x float] into a 32-bit integer.
///
///    If the converted value does not fit in a 32-bit integer, raises a
///    floating-point invalid exception. If the exception is masked, returns
///    the most negative integer.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCVTSS2SI / CVTSS2SI </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the conversion.
/// \returns A 32-bit integer containing the converted value.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_cvtss_si32(__m128 __a)
{
  return __builtin_ia32_cvtss2si((__v4sf)__a);
}

/// Converts a float value contained in the lower 32 bits of a vector of
///    [4 x float] into a 32-bit integer.
///
///    If the converted value does not fit in a 32-bit integer, raises a
///    floating-point invalid exception. If the exception is masked, returns
///    the most negative integer.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCVTSS2SI / CVTSS2SI </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the conversion.
/// \returns A 32-bit integer containing the converted value.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_cvt_ss2si(__m128 __a)
{
  return _mm_cvtss_si32(__a);
}

#ifdef __x86_64__

/// Converts a float value contained in the lower 32 bits of a vector of
///    [4 x float] into a 64-bit integer.
///
///    If the converted value does not fit in a 32-bit integer, raises a
///    floating-point invalid exception. If the exception is masked, returns
///    the most negative integer.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCVTSS2SI / CVTSS2SI </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the conversion.
/// \returns A 64-bit integer containing the converted value.
static __inline__ long long __DEFAULT_FN_ATTRS
_mm_cvtss_si64(__m128 __a)
{
  return __builtin_ia32_cvtss2si64((__v4sf)__a);
}

#endif

/// Converts two low-order float values in a 128-bit vector of
///    [4 x float] into a 64-bit vector of [2 x i32].
///
///    If a converted value does not fit in a 32-bit integer, raises a
///    floating-point invalid exception. If the exception is masked, returns
///    the most negative integer.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPS2PI </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \returns A 64-bit integer vector containing the converted values.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_cvtps_pi32(__m128 __a)
{
  return __trunc64(__builtin_ia32_cvtps2dq((__v4sf)__zeroupper64(__a)));
}

/// Converts two low-order float values in a 128-bit vector of
///    [4 x float] into a 64-bit vector of [2 x i32].
///
///    If a converted value does not fit in a 32-bit integer, raises a
///    floating-point invalid exception. If the exception is masked, returns
///    the most negative integer.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPS2PI </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \returns A 64-bit integer vector containing the converted values.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_cvt_ps2pi(__m128 __a)
{
  return _mm_cvtps_pi32(__a);
}

/// Converts the lower (first) element of a vector of [4 x float] into a signed
///    truncated (rounded toward zero) 32-bit integer.
///
///    If the converted value does not fit in a 32-bit integer, raises a
///    floating-point invalid exception. If the exception is masked, returns
///    the most negative integer.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCVTTSS2SI / CVTTSS2SI </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the conversion.
/// \returns A 32-bit integer containing the converted value.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_cvttss_si32(__m128 __a)
{
  return __builtin_ia32_cvttss2si((__v4sf)__a);
}

/// Converts the lower (first) element of a vector of [4 x float] into a signed
///    truncated (rounded toward zero) 32-bit integer.
///
///    If the converted value does not fit in a 32-bit integer, raises a
///    floating-point invalid exception. If the exception is masked, returns
///    the most negative integer.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCVTTSS2SI / CVTTSS2SI </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the conversion.
/// \returns A 32-bit integer containing the converted value.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_cvtt_ss2si(__m128 __a)
{
  return _mm_cvttss_si32(__a);
}

#ifdef __x86_64__
/// Converts the lower (first) element of a vector of [4 x float] into a signed
///    truncated (rounded toward zero) 64-bit integer.
///
///    If the converted value does not fit in a 64-bit integer, raises a
///    floating-point invalid exception. If the exception is masked, returns
///    the most negative integer.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCVTTSS2SI / CVTTSS2SI </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the conversion.
/// \returns A 64-bit integer containing the converted value.
static __inline__ long long __DEFAULT_FN_ATTRS
_mm_cvttss_si64(__m128 __a)
{
  return __builtin_ia32_cvttss2si64((__v4sf)__a);
}
#endif

/// Converts the lower (first) two elements of a 128-bit vector of [4 x float]
///    into two signed truncated (rounded toward zero) 32-bit integers,
///    returned in a 64-bit vector of [2 x i32].
///
///    If a converted value does not fit in a 32-bit integer, raises a
///    floating-point invalid exception. If the exception is masked, returns
///    the most negative integer.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTTPS2PI / VTTPS2PI </c>
///   instructions.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \returns A 64-bit integer vector containing the converted values.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_cvttps_pi32(__m128 __a)
{
  return __trunc64(__builtin_ia32_cvttps2dq((__v4sf)__zeroupper64(__a)));
}

/// Converts the lower (first) two elements of a 128-bit vector of [4 x float]
///    into two signed truncated (rounded toward zero) 64-bit integers,
///    returned in a 64-bit vector of [2 x i32].
///
///    If a converted value does not fit in a 32-bit integer, raises a
///    floating-point invalid exception. If the exception is masked, returns
///    the most negative integer.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTTPS2PI </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \returns A 64-bit integer vector containing the converted values.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_cvtt_ps2pi(__m128 __a)
{
  return _mm_cvttps_pi32(__a);
}

/// Converts a 32-bit signed integer value into a floating point value
///    and writes it to the lower 32 bits of the destination. The remaining
///    higher order elements of the destination vector are copied from the
///    corresponding elements in the first operand.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCVTSI2SS / CVTSI2SS </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 32-bit signed integer operand containing the value to be converted.
/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
///    converted value of the second operand. The upper 96 bits are copied from
///    the upper 96 bits of the first operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_cvtsi32_ss(__m128 __a,
                                                                     int __b) {
  __a[0] = __b;
  return __a;
}

/// Converts a 32-bit signed integer value into a floating point value
///    and writes it to the lower 32 bits of the destination. The remaining
///    higher order elements of the destination are copied from the
///    corresponding elements in the first operand.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCVTSI2SS / CVTSI2SS </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 32-bit signed integer operand containing the value to be converted.
/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
///    converted value of the second operand. The upper 96 bits are copied from
///    the upper 96 bits of the first operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_cvt_si2ss(__m128 __a,
                                                                    int __b) {
  return _mm_cvtsi32_ss(__a, __b);
}

#ifdef __x86_64__

/// Converts a 64-bit signed integer value into a floating point value
///    and writes it to the lower 32 bits of the destination. The remaining
///    higher order elements of the destination are copied from the
///    corresponding elements in the first operand.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VCVTSI2SS / CVTSI2SS </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 64-bit signed integer operand containing the value to be converted.
/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
///    converted value of the second operand. The upper 96 bits are copied from
///    the upper 96 bits of the first operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_cvtsi64_ss(__m128 __a, long long __b) {
  __a[0] = __b;
  return __a;
}

#endif

/// Converts two elements of a 64-bit vector of [2 x i32] into two
///    floating point values and writes them to the lower 64-bits of the
///    destination. The remaining higher order elements of the destination are
///    copied from the corresponding elements in the first operand.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPI2PS </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 64-bit vector of [2 x i32]. The elements in this vector are converted
///    and written to the corresponding low-order elements in the destination.
/// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
///    converted value of the second operand. The upper 64 bits are copied from
///    the upper 64 bits of the first operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
_mm_cvtpi32_ps(__m128 __a, __m64 __b)
{
  return (__m128)__builtin_shufflevector(
      (__v4sf)__a,
      __builtin_convertvector((__v4si)__zext128(__b), __v4sf),
      4, 5, 2, 3);
}

/// Converts two elements of a 64-bit vector of [2 x i32] into two
///    floating point values and writes them to the lower 64-bits of the
///    destination. The remaining higher order elements of the destination are
///    copied from the corresponding elements in the first operand.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPI2PS </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float].
/// \param __b
///    A 64-bit vector of [2 x i32]. The elements in this vector are converted
///    and written to the corresponding low-order elements in the destination.
/// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
///    converted value from the second operand. The upper 64 bits are copied
///    from the upper 64 bits of the first operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
_mm_cvt_pi2ps(__m128 __a, __m64 __b)
{
  return _mm_cvtpi32_ps(__a, __b);
}

/// Extracts a float value contained in the lower 32 bits of a vector of
///    [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic has no corresponding instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
///    used in the extraction.
/// \returns A 32-bit float containing the extracted value.
static __inline__ float __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_cvtss_f32(__m128 __a) {
  return __a[0];
}

/// Loads two packed float values from the address \a __p into the
///     high-order bits of a 128-bit vector of [4 x float]. The low-order bits
///     are copied from the low-order bits of the first operand.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVHPD / MOVHPD </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float]. Bits [63:0] are written to bits [63:0]
///    of the destination.
/// \param __p
///    A pointer to two packed float values. Bits [63:0] are written to bits
///    [127:64] of the destination.
/// \returns A 128-bit vector of [4 x float] containing the moved values.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_loadh_pi(__m128 __a, const __m64 *__p)
{
  typedef float __mm_loadh_pi_v2f32 __attribute__((__vector_size__(8)));
  struct __mm_loadh_pi_struct {
    __mm_loadh_pi_v2f32 __u;
  } __attribute__((__packed__, __may_alias__));
  __mm_loadh_pi_v2f32 __b = ((const struct __mm_loadh_pi_struct*)__p)->__u;
  __m128 __bb = __builtin_shufflevector(__b, __b, 0, 1, 0, 1);
  return __builtin_shufflevector(__a, __bb, 0, 1, 4, 5);
}

/// Loads two packed float values from the address \a __p into the
///    low-order bits of a 128-bit vector of [4 x float]. The high-order bits
///    are copied from the high-order bits of the first operand.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVLPD / MOVLPD </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float]. Bits [127:64] are written to bits
///    [127:64] of the destination.
/// \param __p
///    A pointer to two packed float values. Bits [63:0] are written to bits
///    [63:0] of the destination.
/// \returns A 128-bit vector of [4 x float] containing the moved values.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_loadl_pi(__m128 __a, const __m64 *__p)
{
  typedef float __mm_loadl_pi_v2f32 __attribute__((__vector_size__(8)));
  struct __mm_loadl_pi_struct {
    __mm_loadl_pi_v2f32 __u;
  } __attribute__((__packed__, __may_alias__));
  __mm_loadl_pi_v2f32 __b = ((const struct __mm_loadl_pi_struct*)__p)->__u;
  __m128 __bb = __builtin_shufflevector(__b, __b, 0, 1, 0, 1);
  return __builtin_shufflevector(__a, __bb, 4, 5, 2, 3);
}

/// Constructs a 128-bit floating-point vector of [4 x float]. The lower
///    32 bits of the vector are initialized with the single-precision
///    floating-point value loaded from a specified memory location. The upper
///    96 bits are set to zero.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVSS / MOVSS </c> instruction.
///
/// \param __p
///    A pointer to a 32-bit memory location containing a single-precision
///    floating-point value.
/// \returns An initialized 128-bit floating-point vector of [4 x float]. The
///    lower 32 bits contain the value loaded from the memory location. The
///    upper 96 bits are set to zero.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_load_ss(const float *__p)
{
  struct __mm_load_ss_struct {
    float __u;
  } __attribute__((__packed__, __may_alias__));
  float __u = ((const struct __mm_load_ss_struct*)__p)->__u;
  return __extension__ (__m128){ __u, 0, 0, 0 };
}

/// Loads a 32-bit float value and duplicates it to all four vector
///    elements of a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VBROADCASTSS / MOVSS + shuffling </c>
///    instruction.
///
/// \param __p
///    A pointer to a float value to be loaded and duplicated.
/// \returns A 128-bit vector of [4 x float] containing the loaded and
///    duplicated values.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_load1_ps(const float *__p)
{
  struct __mm_load1_ps_struct {
    float __u;
  } __attribute__((__packed__, __may_alias__));
  float __u = ((const struct __mm_load1_ps_struct*)__p)->__u;
  return __extension__ (__m128){ __u, __u, __u, __u };
}

#define        _mm_load_ps1(p) _mm_load1_ps(p)

/// Loads a 128-bit floating-point vector of [4 x float] from an aligned
///    memory location.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVAPS / MOVAPS </c> instruction.
///
/// \param __p
///    A pointer to a 128-bit memory location. The address of the memory
///    location has to be 128-bit aligned.
/// \returns A 128-bit vector of [4 x float] containing the loaded values.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_load_ps(const float *__p)
{
  return *(const __m128*)__p;
}

/// Loads a 128-bit floating-point vector of [4 x float] from an
///    unaligned memory location.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVUPS / MOVUPS </c> instruction.
///
/// \param __p
///    A pointer to a 128-bit memory location. The address of the memory
///    location does not have to be aligned.
/// \returns A 128-bit vector of [4 x float] containing the loaded values.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_loadu_ps(const float *__p)
{
  struct __loadu_ps {
    __m128_u __v;
  } __attribute__((__packed__, __may_alias__));
  return ((const struct __loadu_ps*)__p)->__v;
}

/// Loads four packed float values, in reverse order, from an aligned
///    memory location to 32-bit elements in a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVAPS / MOVAPS + shuffling </c>
///    instruction.
///
/// \param __p
///    A pointer to a 128-bit memory location. The address of the memory
///    location has to be 128-bit aligned.
/// \returns A 128-bit vector of [4 x float] containing the moved values, loaded
///    in reverse order.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_loadr_ps(const float *__p)
{
  __m128 __a = _mm_load_ps(__p);
  return __builtin_shufflevector((__v4sf)__a, (__v4sf)__a, 3, 2, 1, 0);
}

/// Create a 128-bit vector of [4 x float] with undefined values.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic has no corresponding instruction.
///
/// \returns A 128-bit vector of [4 x float] containing undefined values.
static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_undefined_ps(void)
{
  return (__m128)__builtin_ia32_undef128();
}

/// Constructs a 128-bit floating-point vector of [4 x float]. The lower
///    32 bits of the vector are initialized with the specified single-precision
///    floating-point value. The upper 96 bits are set to zero.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVSS / MOVSS </c> instruction.
///
/// \param __w
///    A single-precision floating-point value used to initialize the lower 32
///    bits of the result.
/// \returns An initialized 128-bit floating-point vector of [4 x float]. The
///    lower 32 bits contain the value provided in the source operand. The
///    upper 96 bits are set to zero.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_set_ss(float __w) {
  return __extension__ (__m128){ __w, 0.0f, 0.0f, 0.0f };
}

/// Constructs a 128-bit floating-point vector of [4 x float], with each
///    of the four single-precision floating-point vector elements set to the
///    specified single-precision floating-point value.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VPERMILPS / PERMILPS </c> instruction.
///
/// \param __w
///    A single-precision floating-point value used to initialize each vector
///    element of the result.
/// \returns An initialized 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_set1_ps(float __w) {
  return __extension__ (__m128){ __w, __w, __w, __w };
}

/* Microsoft specific. */
/// Constructs a 128-bit floating-point vector of [4 x float], with each
///    of the four single-precision floating-point vector elements set to the
///    specified single-precision floating-point value.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VPERMILPS / PERMILPS </c> instruction.
///
/// \param __w
///    A single-precision floating-point value used to initialize each vector
///    element of the result.
/// \returns An initialized 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_set_ps1(float __w) {
    return _mm_set1_ps(__w);
}

/// Constructs a 128-bit floating-point vector of [4 x float]
///    initialized with the specified single-precision floating-point values.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic is a utility function and does not correspond to a specific
///    instruction.
///
/// \param __z
///    A single-precision floating-point value used to initialize bits [127:96]
///    of the result.
/// \param __y
///    A single-precision floating-point value used to initialize bits [95:64]
///    of the result.
/// \param __x
///    A single-precision floating-point value used to initialize bits [63:32]
///    of the result.
/// \param __w
///    A single-precision floating-point value used to initialize bits [31:0]
///    of the result.
/// \returns An initialized 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_set_ps(float __z, float __y, float __x, float __w) {
  return __extension__ (__m128){ __w, __x, __y, __z };
}

/// Constructs a 128-bit floating-point vector of [4 x float],
///    initialized in reverse order with the specified 32-bit single-precision
///    float-point values.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic is a utility function and does not correspond to a specific
///    instruction.
///
/// \param __z
///    A single-precision floating-point value used to initialize bits [31:0]
///    of the result.
/// \param __y
///    A single-precision floating-point value used to initialize bits [63:32]
///    of the result.
/// \param __x
///    A single-precision floating-point value used to initialize bits [95:64]
///    of the result.
/// \param __w
///    A single-precision floating-point value used to initialize bits [127:96]
///    of the result.
/// \returns An initialized 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_setr_ps(float __z, float __y, float __x, float __w) {
  return __extension__ (__m128){ __z, __y, __x, __w };
}

/// Constructs a 128-bit floating-point vector of [4 x float] initialized
///    to zero.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VXORPS / XORPS </c> instruction.
///
/// \returns An initialized 128-bit floating-point vector of [4 x float] with
///    all elements set to zero.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_setzero_ps(void) {
  return __extension__ (__m128){ 0.0f, 0.0f, 0.0f, 0.0f };
}

/// Stores the upper 64 bits of a 128-bit vector of [4 x float] to a
///    memory location.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VPEXTRQ / PEXTRQ </c> instruction.
///
/// \param __p
///    A pointer to a 64-bit memory location.
/// \param __a
///    A 128-bit vector of [4 x float] containing the values to be stored.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_storeh_pi(__m64 *__p, __m128 __a)
{
  typedef float __mm_storeh_pi_v2f32 __attribute__((__vector_size__(8)));
  struct __mm_storeh_pi_struct {
    __mm_storeh_pi_v2f32 __u;
  } __attribute__((__packed__, __may_alias__));
  ((struct __mm_storeh_pi_struct*)__p)->__u = __builtin_shufflevector(__a, __a, 2, 3);
}

/// Stores the lower 64 bits of a 128-bit vector of [4 x float] to a
///     memory location.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVLPS / MOVLPS </c> instruction.
///
/// \param __p
///    A pointer to a memory location that will receive the float values.
/// \param __a
///    A 128-bit vector of [4 x float] containing the values to be stored.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_storel_pi(__m64 *__p, __m128 __a)
{
  typedef float __mm_storeh_pi_v2f32 __attribute__((__vector_size__(8)));
  struct __mm_storeh_pi_struct {
    __mm_storeh_pi_v2f32 __u;
  } __attribute__((__packed__, __may_alias__));
  ((struct __mm_storeh_pi_struct*)__p)->__u = __builtin_shufflevector(__a, __a, 0, 1);
}

/// Stores the lower 32 bits of a 128-bit vector of [4 x float] to a
///     memory location.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVSS / MOVSS </c> instruction.
///
/// \param __p
///    A pointer to a 32-bit memory location.
/// \param __a
///    A 128-bit vector of [4 x float] containing the value to be stored.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_store_ss(float *__p, __m128 __a)
{
  struct __mm_store_ss_struct {
    float __u;
  } __attribute__((__packed__, __may_alias__));
  ((struct __mm_store_ss_struct*)__p)->__u = __a[0];
}

/// Stores a 128-bit vector of [4 x float] to an unaligned memory
///    location.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVUPS / MOVUPS </c> instruction.
///
/// \param __p
///    A pointer to a 128-bit memory location. The address of the memory
///    location does not have to be aligned.
/// \param __a
///    A 128-bit vector of [4 x float] containing the values to be stored.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_storeu_ps(float *__p, __m128 __a)
{
  struct __storeu_ps {
    __m128_u __v;
  } __attribute__((__packed__, __may_alias__));
  ((struct __storeu_ps*)__p)->__v = __a;
}

/// Stores a 128-bit vector of [4 x float] into an aligned memory
///    location.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVAPS / MOVAPS </c> instruction.
///
/// \param __p
///    A pointer to a 128-bit memory location. The address of the memory
///    location has to be 16-byte aligned.
/// \param __a
///    A 128-bit vector of [4 x float] containing the values to be stored.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_store_ps(float *__p, __m128 __a)
{
  *(__m128*)__p = __a;
}

/// Stores the lower 32 bits of a 128-bit vector of [4 x float] into
///    four contiguous elements in an aligned memory location.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to <c> VMOVAPS / MOVAPS + shuffling </c>
///    instruction.
///
/// \param __p
///    A pointer to a 128-bit memory location.
/// \param __a
///    A 128-bit vector of [4 x float] whose lower 32 bits are stored to each
///    of the four contiguous elements pointed by \a __p.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_store1_ps(float *__p, __m128 __a)
{
  __a = __builtin_shufflevector((__v4sf)__a, (__v4sf)__a, 0, 0, 0, 0);
  _mm_store_ps(__p, __a);
}

/// Stores the lower 32 bits of a 128-bit vector of [4 x float] into
///    four contiguous elements in an aligned memory location.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to <c> VMOVAPS / MOVAPS + shuffling </c>
///    instruction.
///
/// \param __p
///    A pointer to a 128-bit memory location.
/// \param __a
///    A 128-bit vector of [4 x float] whose lower 32 bits are stored to each
///    of the four contiguous elements pointed by \a __p.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_store_ps1(float *__p, __m128 __a)
{
  _mm_store1_ps(__p, __a);
}

/// Stores float values from a 128-bit vector of [4 x float] to an
///    aligned memory location in reverse order.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVAPS / MOVAPS + shuffling </c>
///    instruction.
///
/// \param __p
///    A pointer to a 128-bit memory location. The address of the memory
///    location has to be 128-bit aligned.
/// \param __a
///    A 128-bit vector of [4 x float] containing the values to be stored.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_storer_ps(float *__p, __m128 __a)
{
  __a = __builtin_shufflevector((__v4sf)__a, (__v4sf)__a, 3, 2, 1, 0);
  _mm_store_ps(__p, __a);
}

#define _MM_HINT_ET0 7
#define _MM_HINT_ET1 6
#define _MM_HINT_T0  3
#define _MM_HINT_T1  2
#define _MM_HINT_T2  1
#define _MM_HINT_NTA 0

#ifndef _MSC_VER
/* FIXME: We have to #define this because "sel" must be a constant integer, and
   Sema doesn't do any form of constant propagation yet. */

/// Loads one cache line of data from the specified address to a location
///    closer to the processor.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// void _mm_prefetch(const void *a, const int sel);
/// \endcode
///
/// This intrinsic corresponds to the <c> PREFETCHNTA </c> instruction.
///
/// \param a
///    A pointer to a memory location containing a cache line of data.
/// \param sel
///    A predefined integer constant specifying the type of prefetch
///    operation: \n
///    _MM_HINT_NTA: Move data using the non-temporal access (NTA) hint. The
///    PREFETCHNTA instruction will be generated. \n
///    _MM_HINT_T0: Move data using the T0 hint. The PREFETCHT0 instruction will
///    be generated. \n
///    _MM_HINT_T1: Move data using the T1 hint. The PREFETCHT1 instruction will
///    be generated. \n
///    _MM_HINT_T2: Move data using the T2 hint. The PREFETCHT2 instruction will
///    be generated.
#define _mm_prefetch(a, sel) (__builtin_prefetch((const void *)(a), \
                                                 ((sel) >> 2) & 1, (sel) & 0x3))
#endif

/// Stores a 64-bit integer in the specified aligned memory location. To
///    minimize caching, the data is flagged as non-temporal (unlikely to be
///    used again soon).
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> MOVNTQ </c> instruction.
///
/// \param __p
///    A pointer to an aligned memory location used to store the register value.
/// \param __a
///    A 64-bit integer containing the value to be stored.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_stream_pi(void *__p, __m64 __a)
{
  __builtin_nontemporal_store(__a, (__m64 *)__p);
}

/// Moves packed float values from a 128-bit vector of [4 x float] to a
///    128-bit aligned memory location. To minimize caching, the data is flagged
///    as non-temporal (unlikely to be used again soon).
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVNTPS / MOVNTPS </c> instruction.
///
/// \param __p
///    A pointer to a 128-bit aligned memory location that will receive the
///    single-precision floating-point values.
/// \param __a
///    A 128-bit vector of [4 x float] containing the values to be moved.
static __inline__ void __DEFAULT_FN_ATTRS
_mm_stream_ps(void *__p, __m128 __a)
{
  __builtin_nontemporal_store((__v4sf)__a, (__v4sf*)__p);
}

#if defined(__cplusplus)
extern "C" {
#endif

/// Forces strong memory ordering (serialization) between store
///    instructions preceding this instruction and store instructions following
///    this instruction, ensuring the system completes all previous stores
///    before executing subsequent stores.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> SFENCE </c> instruction.
///
void _mm_sfence(void);

#if defined(__cplusplus)
} // extern "C"
#endif

/// Extracts 16-bit element from a 64-bit vector of [4 x i16] and
///    returns it, as specified by the immediate integer operand.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// int _mm_extract_pi16(__m64 a, int n);
/// \endcode
///
/// This intrinsic corresponds to the <c> VPEXTRW / PEXTRW </c> instruction.
///
/// \param a
///    A 64-bit vector of [4 x i16].
/// \param n
///    An immediate integer operand that determines which bits are extracted: \n
///    0: Bits [15:0] are copied to the destination. \n
///    1: Bits [31:16] are copied to the destination. \n
///    2: Bits [47:32] are copied to the destination. \n
///    3: Bits [63:48] are copied to the destination.
/// \returns A 16-bit integer containing the extracted 16 bits of packed data.
#define _mm_extract_pi16(a, n) \
  ((int)(unsigned short)__builtin_ia32_vec_ext_v4hi((__v4hi)a, (int)n))

/// Copies data from the 64-bit vector of [4 x i16] to the destination,
///    and inserts the lower 16-bits of an integer operand at the 16-bit offset
///    specified by the immediate operand \a n.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// __m64 _mm_insert_pi16(__m64 a, int d, int n);
/// \endcode
///
/// This intrinsic corresponds to the <c> PINSRW </c> instruction.
///
/// \param a
///    A 64-bit vector of [4 x i16].
/// \param d
///    An integer. The lower 16-bit value from this operand is written to the
///    destination at the offset specified by operand \a n.
/// \param n
///    An immediate integer operant that determines which the bits to be used
///    in the destination. \n
///    0: Bits [15:0] are copied to the destination. \n
///    1: Bits [31:16] are copied to the destination. \n
///    2: Bits [47:32] are copied to the destination. \n
///    3: Bits [63:48] are copied to the destination.  \n
///    The remaining bits in the destination are copied from the corresponding
///    bits in operand \a a.
/// \returns A 64-bit integer vector containing the copied packed data from the
///    operands.
#define _mm_insert_pi16(a, d, n) \
  ((__m64)__builtin_ia32_vec_set_v4hi((__v4hi)a, (int)d, (int)n))

/// Compares each of the corresponding packed 16-bit integer values of
///    the 64-bit integer vectors, and writes the greater value to the
///    corresponding bits in the destination.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> PMAXSW </c> instruction.
///
/// \param __a
///    A 64-bit integer vector containing one of the source operands.
/// \param __b
///    A 64-bit integer vector containing one of the source operands.
/// \returns A 64-bit integer vector containing the comparison results.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_max_pi16(__m64 __a, __m64 __b)
{
  return (__m64)__builtin_elementwise_max((__v4hi)__a, (__v4hi)__b);
}

/// Compares each of the corresponding packed 8-bit unsigned integer
///    values of the 64-bit integer vectors, and writes the greater value to the
///    corresponding bits in the destination.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> PMAXUB </c> instruction.
///
/// \param __a
///    A 64-bit integer vector containing one of the source operands.
/// \param __b
///    A 64-bit integer vector containing one of the source operands.
/// \returns A 64-bit integer vector containing the comparison results.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_max_pu8(__m64 __a, __m64 __b)
{
  return (__m64)__builtin_elementwise_max((__v8qu)__a, (__v8qu)__b);
}

/// Compares each of the corresponding packed 16-bit integer values of
///    the 64-bit integer vectors, and writes the lesser value to the
///    corresponding bits in the destination.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> PMINSW </c> instruction.
///
/// \param __a
///    A 64-bit integer vector containing one of the source operands.
/// \param __b
///    A 64-bit integer vector containing one of the source operands.
/// \returns A 64-bit integer vector containing the comparison results.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_min_pi16(__m64 __a, __m64 __b)
{
  return (__m64)__builtin_elementwise_min((__v4hi)__a, (__v4hi)__b);
}

/// Compares each of the corresponding packed 8-bit unsigned integer
///    values of the 64-bit integer vectors, and writes the lesser value to the
///    corresponding bits in the destination.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> PMINUB </c> instruction.
///
/// \param __a
///    A 64-bit integer vector containing one of the source operands.
/// \param __b
///    A 64-bit integer vector containing one of the source operands.
/// \returns A 64-bit integer vector containing the comparison results.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_min_pu8(__m64 __a, __m64 __b)
{
  return (__m64)__builtin_elementwise_min((__v8qu)__a, (__v8qu)__b);
}

/// Takes the most significant bit from each 8-bit element in a 64-bit
///    integer vector to create an 8-bit mask value. Zero-extends the value to
///    32-bit integer and writes it to the destination.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> PMOVMSKB </c> instruction.
///
/// \param __a
///    A 64-bit integer vector containing the values with bits to be extracted.
/// \returns The most significant bit from each 8-bit element in \a __a,
///    written to bits [7:0].
static __inline__ int __DEFAULT_FN_ATTRS_SSE2
_mm_movemask_pi8(__m64 __a)
{
  return __builtin_ia32_pmovmskb128((__v16qi)__zext128(__a));
}

/// Multiplies packed 16-bit unsigned integer values and writes the
///    high-order 16 bits of each 32-bit product to the corresponding bits in
///    the destination.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> PMULHUW </c> instruction.
///
/// \param __a
///    A 64-bit integer vector containing one of the source operands.
/// \param __b
///    A 64-bit integer vector containing one of the source operands.
/// \returns A 64-bit integer vector containing the products of both operands.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_mulhi_pu16(__m64 __a, __m64 __b)
{
  return __trunc64(__builtin_ia32_pmulhuw128((__v8hi)__anyext128(__a),
                                             (__v8hi)__anyext128(__b)));
}

/// Shuffles the 4 16-bit integers from a 64-bit integer vector to the
///    destination, as specified by the immediate value operand.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// __m64 _mm_shuffle_pi16(__m64 a, const int n);
/// \endcode
///
/// This intrinsic corresponds to the <c> PSHUFW </c> instruction.
///
/// \param a
///    A 64-bit integer vector containing the values to be shuffled.
/// \param n
///    An immediate value containing an 8-bit value specifying which elements to
///    copy from \a a. The destinations within the 64-bit destination are
///    assigned values as follows: \n
///    Bits [1:0] are used to assign values to bits [15:0] in the
///    destination. \n
///    Bits [3:2] are used to assign values to bits [31:16] in the
///    destination. \n
///    Bits [5:4] are used to assign values to bits [47:32] in the
///    destination. \n
///    Bits [7:6] are used to assign values to bits [63:48] in the
///    destination. \n
///    Bit value assignments: \n
///    00: assigned from bits [15:0] of \a a. \n
///    01: assigned from bits [31:16] of \a a. \n
///    10: assigned from bits [47:32] of \a a. \n
///    11: assigned from bits [63:48] of \a a. \n
///    Note: To generate a mask, you can use the \c _MM_SHUFFLE macro.
///    <c>_MM_SHUFFLE(b6, b4, b2, b0)</c> can create an 8-bit mask of the form
///    <c>[b6, b4, b2, b0]</c>.
/// \returns A 64-bit integer vector containing the shuffled values.
#define _mm_shuffle_pi16(a, n)                                                 \
  ((__m64)__builtin_shufflevector((__v4hi)(__m64)(a), __extension__(__v4hi){}, \
                                  (n) & 0x3, ((n) >> 2) & 0x3,                 \
                                  ((n) >> 4) & 0x3, ((n) >> 6) & 0x3))

/// Conditionally copies the values from each 8-bit element in the first
///    64-bit integer vector operand to the specified memory location, as
///    specified by the most significant bit in the corresponding element in the
///    second 64-bit integer vector operand.
///
///    To minimize caching, the data is flagged as non-temporal
///    (unlikely to be used again soon).
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> MASKMOVQ </c> instruction.
///
/// \param __d
///    A 64-bit integer vector containing the values with elements to be copied.
/// \param __n
///    A 64-bit integer vector operand. The most significant bit from each 8-bit
///    element determines whether the corresponding element in operand \a __d
///    is copied. If the most significant bit of a given element is 1, the
///    corresponding element in operand \a __d is copied.
/// \param __p
///    A pointer to a 64-bit memory location that will receive the conditionally
///    copied integer values. The address of the memory location does not have
///    to be aligned.
static __inline__ void __DEFAULT_FN_ATTRS_SSE2
_mm_maskmove_si64(__m64 __d, __m64 __n, char *__p)
{
  // This is complex, because we need to support the case where __p is pointing
  // within the last 15 to 8 bytes of a page. In that case, using a 128-bit
  // write might cause a trap where a 64-bit maskmovq would not. (Memory
  // locations not selected by the mask bits might still cause traps.)
  __m128i __d128  = __anyext128(__d);
  __m128i __n128  = __zext128(__n);
  if (((__SIZE_TYPE__)__p & 0xfff) >= 4096-15 &&
      ((__SIZE_TYPE__)__p & 0xfff) <= 4096-8) {
    // If there's a risk of spurious trap due to a 128-bit write, back up the
    // pointer by 8 bytes and shift values in registers to match.
    __p -= 8;
    __d128 = __builtin_ia32_pslldqi128_byteshift((__v2di)__d128, 8);
    __n128 = __builtin_ia32_pslldqi128_byteshift((__v2di)__n128, 8);
  }

  __builtin_ia32_maskmovdqu((__v16qi)__d128, (__v16qi)__n128, __p);
}

/// Computes the rounded averages of the packed unsigned 8-bit integer
///    values and writes the averages to the corresponding bits in the
///    destination.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> PAVGB </c> instruction.
///
/// \param __a
///    A 64-bit integer vector containing one of the source operands.
/// \param __b
///    A 64-bit integer vector containing one of the source operands.
/// \returns A 64-bit integer vector containing the averages of both operands.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_avg_pu8(__m64 __a, __m64 __b)
{
  return __trunc64(__builtin_ia32_pavgb128((__v16qi)__anyext128(__a),
                                           (__v16qi)__anyext128(__b)));
}

/// Computes the rounded averages of the packed unsigned 16-bit integer
///    values and writes the averages to the corresponding bits in the
///    destination.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> PAVGW </c> instruction.
///
/// \param __a
///    A 64-bit integer vector containing one of the source operands.
/// \param __b
///    A 64-bit integer vector containing one of the source operands.
/// \returns A 64-bit integer vector containing the averages of both operands.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_avg_pu16(__m64 __a, __m64 __b)
{
  return __trunc64(__builtin_ia32_pavgw128((__v8hi)__anyext128(__a),
                                           (__v8hi)__anyext128(__b)));
}

/// Subtracts the corresponding 8-bit unsigned integer values of the two
///    64-bit vector operands and computes the absolute value for each of the
///    difference. Then sum of the 8 absolute differences is written to the
///    bits [15:0] of the destination; the remaining bits [63:16] are cleared.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> PSADBW </c> instruction.
///
/// \param __a
///    A 64-bit integer vector containing one of the source operands.
/// \param __b
///    A 64-bit integer vector containing one of the source operands.
/// \returns A 64-bit integer vector whose lower 16 bits contain the sums of the
///    sets of absolute differences between both operands. The upper bits are
///    cleared.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_sad_pu8(__m64 __a, __m64 __b)
{
  return __trunc64(__builtin_ia32_psadbw128((__v16qi)__zext128(__a),
                                            (__v16qi)__zext128(__b)));
}

#if defined(__cplusplus)
extern "C" {
#endif

/// Returns the contents of the MXCSR register as a 32-bit unsigned
///    integer value.
///
///    There are several groups of macros associated with this
///    intrinsic, including:
///    <ul>
///    <li>
///      For checking exception states: _MM_EXCEPT_INVALID, _MM_EXCEPT_DIV_ZERO,
///      _MM_EXCEPT_DENORM, _MM_EXCEPT_OVERFLOW, _MM_EXCEPT_UNDERFLOW,
///      _MM_EXCEPT_INEXACT. There is a convenience wrapper
///      _MM_GET_EXCEPTION_STATE().
///    </li>
///    <li>
///      For checking exception masks: _MM_MASK_UNDERFLOW, _MM_MASK_OVERFLOW,
///      _MM_MASK_INVALID, _MM_MASK_DENORM, _MM_MASK_DIV_ZERO, _MM_MASK_INEXACT.
///      There is a convenience wrapper _MM_GET_EXCEPTION_MASK().
///    </li>
///    <li>
///      For checking rounding modes: _MM_ROUND_NEAREST, _MM_ROUND_DOWN,
///      _MM_ROUND_UP, _MM_ROUND_TOWARD_ZERO. There is a convenience wrapper
///      _MM_GET_ROUNDING_MODE().
///    </li>
///    <li>
///      For checking flush-to-zero mode: _MM_FLUSH_ZERO_ON, _MM_FLUSH_ZERO_OFF.
///      There is a convenience wrapper _MM_GET_FLUSH_ZERO_MODE().
///    </li>
///    <li>
///      For checking denormals-are-zero mode: _MM_DENORMALS_ZERO_ON,
///      _MM_DENORMALS_ZERO_OFF. There is a convenience wrapper
///      _MM_GET_DENORMALS_ZERO_MODE().
///    </li>
///    </ul>
///
///    For example, the following expression checks if an overflow exception has
///    occurred:
///    \code
///      ( _mm_getcsr() & _MM_EXCEPT_OVERFLOW )
///    \endcode
///
///    The following expression gets the current rounding mode:
///    \code
///      _MM_GET_ROUNDING_MODE()
///    \endcode
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VSTMXCSR / STMXCSR </c> instruction.
///
/// \returns A 32-bit unsigned integer containing the contents of the MXCSR
///    register.
unsigned int _mm_getcsr(void);

/// Sets the MXCSR register with the 32-bit unsigned integer value.
///
///    There are several groups of macros associated with this intrinsic,
///    including:
///    <ul>
///    <li>
///      For setting exception states: _MM_EXCEPT_INVALID, _MM_EXCEPT_DIV_ZERO,
///      _MM_EXCEPT_DENORM, _MM_EXCEPT_OVERFLOW, _MM_EXCEPT_UNDERFLOW,
///      _MM_EXCEPT_INEXACT. There is a convenience wrapper
///      _MM_SET_EXCEPTION_STATE(x) where x is one of these macros.
///    </li>
///    <li>
///      For setting exception masks: _MM_MASK_UNDERFLOW, _MM_MASK_OVERFLOW,
///      _MM_MASK_INVALID, _MM_MASK_DENORM, _MM_MASK_DIV_ZERO, _MM_MASK_INEXACT.
///      There is a convenience wrapper _MM_SET_EXCEPTION_MASK(x) where x is one
///      of these macros.
///    </li>
///    <li>
///      For setting rounding modes: _MM_ROUND_NEAREST, _MM_ROUND_DOWN,
///      _MM_ROUND_UP, _MM_ROUND_TOWARD_ZERO. There is a convenience wrapper
///      _MM_SET_ROUNDING_MODE(x) where x is one of these macros.
///    </li>
///    <li>
///      For setting flush-to-zero mode: _MM_FLUSH_ZERO_ON, _MM_FLUSH_ZERO_OFF.
///      There is a convenience wrapper _MM_SET_FLUSH_ZERO_MODE(x) where x is
///      one of these macros.
///    </li>
///    <li>
///      For setting denormals-are-zero mode: _MM_DENORMALS_ZERO_ON,
///      _MM_DENORMALS_ZERO_OFF. There is a convenience wrapper
///      _MM_SET_DENORMALS_ZERO_MODE(x) where x is one of these macros.
///    </li>
///    </ul>
///
///    For example, the following expression causes subsequent floating-point
///    operations to round up:
///      _mm_setcsr(_mm_getcsr() | _MM_ROUND_UP)
///
///    The following example sets the DAZ and FTZ flags:
///    \code
///    void setFlags() {
///      _MM_SET_FLUSH_ZERO_MODE(_MM_FLUSH_ZERO_ON);
///      _MM_SET_DENORMALS_ZERO_MODE(_MM_DENORMALS_ZERO_ON);
///    }
///    \endcode
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VLDMXCSR / LDMXCSR </c> instruction.
///
/// \param __i
///    A 32-bit unsigned integer value to be written to the MXCSR register.
void _mm_setcsr(unsigned int __i);

#if defined(__cplusplus)
} // extern "C"
#endif

/// Selects 4 float values from the 128-bit operands of [4 x float], as
///    specified by the immediate value operand.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// __m128 _mm_shuffle_ps(__m128 a, __m128 b, const int mask);
/// \endcode
///
/// This intrinsic corresponds to the <c> VSHUFPS / SHUFPS </c> instruction.
///
/// \param a
///    A 128-bit vector of [4 x float].
/// \param b
///    A 128-bit vector of [4 x float].
/// \param mask
///    An immediate value containing an 8-bit value specifying which elements to
///    copy from \a a and \a b. \n
///    Bits [3:0] specify the values copied from operand \a a. \n
///    Bits [7:4] specify the values copied from operand \a b. \n
///    The destinations within the 128-bit destination are assigned values as
///    follows: \n
///    Bits [1:0] are used to assign values to bits [31:0] in the
///    destination. \n
///    Bits [3:2] are used to assign values to bits [63:32] in the
///    destination. \n
///    Bits [5:4] are used to assign values to bits [95:64] in the
///    destination. \n
///    Bits [7:6] are used to assign values to bits [127:96] in the
///    destination. \n
///    Bit value assignments: \n
///    00: Bits [31:0] copied from the specified operand. \n
///    01: Bits [63:32] copied from the specified operand. \n
///    10: Bits [95:64] copied from the specified operand. \n
///    11: Bits [127:96] copied from the specified operand. \n
///    Note: To generate a mask, you can use the \c _MM_SHUFFLE macro.
///    <c>_MM_SHUFFLE(b6, b4, b2, b0)</c> can create an 8-bit mask of the form
///    <c>[b6, b4, b2, b0]</c>.
/// \returns A 128-bit vector of [4 x float] containing the shuffled values.
#define _mm_shuffle_ps(a, b, mask) \
  ((__m128)__builtin_ia32_shufps((__v4sf)(__m128)(a), (__v4sf)(__m128)(b), \
                                 (int)(mask)))

/// Unpacks the high-order (index 2,3) values from two 128-bit vectors of
///    [4 x float] and interleaves them into a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VUNPCKHPS / UNPCKHPS </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float]. \n
///    Bits [95:64] are written to bits [31:0] of the destination. \n
///    Bits [127:96] are written to bits [95:64] of the destination.
/// \param __b
///    A 128-bit vector of [4 x float].
///    Bits [95:64] are written to bits [63:32] of the destination. \n
///    Bits [127:96] are written to bits [127:96] of the destination.
/// \returns A 128-bit vector of [4 x float] containing the interleaved values.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_unpackhi_ps(__m128 __a, __m128 __b) {
  return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 2, 6, 3, 7);
}

/// Unpacks the low-order (index 0,1) values from two 128-bit vectors of
///    [4 x float] and interleaves them into a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VUNPCKLPS / UNPCKLPS </c> instruction.
///
/// \param __a
///    A 128-bit vector of [4 x float]. \n
///    Bits [31:0] are written to bits [31:0] of the destination.  \n
///    Bits [63:32] are written to bits [95:64] of the destination.
/// \param __b
///    A 128-bit vector of [4 x float]. \n
///    Bits [31:0] are written to bits [63:32] of the destination. \n
///    Bits [63:32] are written to bits [127:96] of the destination.
/// \returns A 128-bit vector of [4 x float] containing the interleaved values.
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_unpacklo_ps(__m128 __a, __m128 __b) {
  return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 0, 4, 1, 5);
}

/// Constructs a 128-bit floating-point vector of [4 x float]. The lower
///    32 bits are set to the lower 32 bits of the second parameter. The upper
///    96 bits are set to the upper 96 bits of the first parameter.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VBLENDPS / BLENDPS / MOVSS </c>
///    instruction.
///
/// \param __a
///    A 128-bit floating-point vector of [4 x float]. The upper 96 bits are
///    written to the upper 96 bits of the result.
/// \param __b
///    A 128-bit floating-point vector of [4 x float]. The lower 32 bits are
///    written to the lower 32 bits of the result.
/// \returns A 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_move_ss(__m128 __a, __m128 __b) {
  __a[0] = __b[0];
  return __a;
}

/// Constructs a 128-bit floating-point vector of [4 x float]. The lower
///    64 bits are set to the upper 64 bits of the second parameter. The upper
///    64 bits are set to the upper 64 bits of the first parameter.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VUNPCKHPD / UNPCKHPD </c> instruction.
///
/// \param __a
///    A 128-bit floating-point vector of [4 x float]. The upper 64 bits are
///    written to the upper 64 bits of the result.
/// \param __b
///    A 128-bit floating-point vector of [4 x float]. The upper 64 bits are
///    written to the lower 64 bits of the result.
/// \returns A 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_movehl_ps(__m128 __a, __m128 __b) {
  return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 6, 7, 2, 3);
}

/// Constructs a 128-bit floating-point vector of [4 x float]. The lower
///    64 bits are set to the lower 64 bits of the first parameter. The upper
///    64 bits are set to the lower 64 bits of the second parameter.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VUNPCKLPD / UNPCKLPD </c> instruction.
///
/// \param __a
///    A 128-bit floating-point vector of [4 x float]. The lower 64 bits are
///    written to the lower 64 bits of the result.
/// \param __b
///    A 128-bit floating-point vector of [4 x float]. The lower 64 bits are
///    written to the upper 64 bits of the result.
/// \returns A 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
_mm_movelh_ps(__m128 __a, __m128 __b) {
  return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 0, 1, 4, 5);
}

/// Converts a 64-bit vector of [4 x i16] into a 128-bit vector of [4 x
///    float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPI2PS + COMPOSITE </c> instruction.
///
/// \param __a
///    A 64-bit vector of [4 x i16]. The elements of the destination are copied
///    from the corresponding elements in this operand.
/// \returns A 128-bit vector of [4 x float] containing the copied and converted
///    values from the operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
_mm_cvtpi16_ps(__m64 __a)
{
  return __builtin_convertvector((__v4hi)__a, __v4sf);
}

/// Converts a 64-bit vector of 16-bit unsigned integer values into a
///    128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPI2PS + COMPOSITE </c> instruction.
///
/// \param __a
///    A 64-bit vector of 16-bit unsigned integer values. The elements of the
///    destination are copied from the corresponding elements in this operand.
/// \returns A 128-bit vector of [4 x float] containing the copied and converted
///    values from the operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
_mm_cvtpu16_ps(__m64 __a)
{
  return __builtin_convertvector((__v4hu)__a, __v4sf);
}

/// Converts the lower four 8-bit values from a 64-bit vector of [8 x i8]
///    into a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPI2PS + COMPOSITE </c> instruction.
///
/// \param __a
///    A 64-bit vector of [8 x i8]. The elements of the destination are copied
///    from the corresponding lower 4 elements in this operand.
/// \returns A 128-bit vector of [4 x float] containing the copied and converted
///    values from the operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
_mm_cvtpi8_ps(__m64 __a)
{
  return __builtin_convertvector(
      __builtin_shufflevector((__v8qs)__a, __extension__ (__v8qs){},
                              0, 1, 2, 3), __v4sf);
}

/// Converts the lower four unsigned 8-bit integer values from a 64-bit
///    vector of [8 x u8] into a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPI2PS + COMPOSITE </c> instruction.
///
/// \param __a
///    A 64-bit vector of unsigned 8-bit integer values. The elements of the
///    destination are copied from the corresponding lower 4 elements in this
///    operand.
/// \returns A 128-bit vector of [4 x float] containing the copied and converted
///    values from the source operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
_mm_cvtpu8_ps(__m64 __a)
{
  return __builtin_convertvector(
      __builtin_shufflevector((__v8qu)__a, __extension__ (__v8qu){},
                              0, 1, 2, 3), __v4sf);
}

/// Converts the two 32-bit signed integer values from each 64-bit vector
///    operand of [2 x i32] into a 128-bit vector of [4 x float].
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPI2PS + COMPOSITE </c> instruction.
///
/// \param __a
///    A 64-bit vector of [2 x i32]. The lower elements of the destination are
///    copied from the elements in this operand.
/// \param __b
///    A 64-bit vector of [2 x i32]. The upper elements of the destination are
///    copied from the elements in this operand.
/// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
///    copied and converted values from the first operand. The upper 64 bits
///    contain the copied and converted values from the second operand.
static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
_mm_cvtpi32x2_ps(__m64 __a, __m64 __b)
{
  return __builtin_convertvector(
      __builtin_shufflevector((__v2si)__a, (__v2si)__b,
                              0, 1, 2, 3), __v4sf);
}

/// Converts each single-precision floating-point element of a 128-bit
///    floating-point vector of [4 x float] into a 16-bit signed integer, and
///    packs the results into a 64-bit integer vector of [4 x i16].
///
///    If the floating-point element is NaN or infinity, or if the
///    floating-point element is greater than 0x7FFFFFFF or less than -0x8000,
///    it is converted to 0x8000. Otherwise if the floating-point element is
///    greater than 0x7FFF, it is converted to 0x7FFF.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPS2PI + COMPOSITE </c> instruction.
///
/// \param __a
///    A 128-bit floating-point vector of [4 x float].
/// \returns A 64-bit integer vector of [4 x i16] containing the converted
///    values.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_cvtps_pi16(__m128 __a)
{
  return __trunc64(__builtin_ia32_packssdw128(
      (__v4si)__builtin_ia32_cvtps2dq((__v4sf)__a), (__v4si)_mm_setzero_ps()));
}

/// Converts each single-precision floating-point element of a 128-bit
///    floating-point vector of [4 x float] into an 8-bit signed integer, and
///    packs the results into the lower 32 bits of a 64-bit integer vector of
///    [8 x i8]. The upper 32 bits of the vector are set to 0.
///
///    If the floating-point element is NaN or infinity, or if the
///    floating-point element is greater than 0x7FFFFFFF or less than -0x80, it
///    is converted to 0x80. Otherwise if the floating-point element is greater
///    than 0x7F, it is converted to 0x7F.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> CVTPS2PI + COMPOSITE </c> instruction.
///
/// \param __a
///    128-bit floating-point vector of [4 x float].
/// \returns A 64-bit integer vector of [8 x i8]. The lower 32 bits contain the
///    converted values and the uppper 32 bits are set to zero.
static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
_mm_cvtps_pi8(__m128 __a)
{
  __m64 __b, __c;

  __b = _mm_cvtps_pi16(__a);
  __c = _mm_setzero_si64();

  return _mm_packs_pi16(__b, __c);
}

/// Extracts the sign bits from each single-precision floating-point
///    element of a 128-bit floating-point vector of [4 x float] and returns the
///    sign bits in bits [0:3] of the result. Bits [31:4] of the result are set
///    to zero.
///
/// \headerfile <x86intrin.h>
///
/// This intrinsic corresponds to the <c> VMOVMSKPS / MOVMSKPS </c> instruction.
///
/// \param __a
///    A 128-bit floating-point vector of [4 x float].
/// \returns A 32-bit integer value. Bits [3:0] contain the sign bits from each
///    single-precision floating-point element of the parameter. Bits [31:4] are
///    set to zero.
static __inline__ int __DEFAULT_FN_ATTRS
_mm_movemask_ps(__m128 __a)
{
  return __builtin_ia32_movmskps((__v4sf)__a);
}

/* Compare */
#define _CMP_EQ_OQ    0x00 /* Equal (ordered, non-signaling)  */
#define _CMP_LT_OS    0x01 /* Less-than (ordered, signaling)  */
#define _CMP_LE_OS    0x02 /* Less-than-or-equal (ordered, signaling)  */
#define _CMP_UNORD_Q  0x03 /* Unordered (non-signaling)  */
#define _CMP_NEQ_UQ   0x04 /* Not-equal (unordered, non-signaling)  */
#define _CMP_NLT_US   0x05 /* Not-less-than (unordered, signaling)  */
#define _CMP_NLE_US   0x06 /* Not-less-than-or-equal (unordered, signaling)  */
#define _CMP_ORD_Q    0x07 /* Ordered (non-signaling)   */

/// Compares each of the corresponding values of two 128-bit vectors of
///    [4 x float], using the operation specified by the immediate integer
///    operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true.
///    If either value in a comparison is NaN, comparisons that are ordered
///    return false, and comparisons that are unordered return true.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// __m128 _mm_cmp_ps(__m128 a, __m128 b, const int c);
/// \endcode
///
/// This intrinsic corresponds to the <c> (V)CMPPS </c> instruction.
///
/// \param a
///    A 128-bit vector of [4 x float].
/// \param b
///    A 128-bit vector of [4 x float].
/// \param c
///    An immediate integer operand, with bits [4:0] specifying which comparison
///    operation to use: \n
///    0x00: Equal (ordered, non-signaling) \n
///    0x01: Less-than (ordered, signaling) \n
///    0x02: Less-than-or-equal (ordered, signaling) \n
///    0x03: Unordered (non-signaling) \n
///    0x04: Not-equal (unordered, non-signaling) \n
///    0x05: Not-less-than (unordered, signaling) \n
///    0x06: Not-less-than-or-equal (unordered, signaling) \n
///    0x07: Ordered (non-signaling) \n
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
#define _mm_cmp_ps(a, b, c)                                                    \
  ((__m128)__builtin_ia32_cmpps((__v4sf)(__m128)(a), (__v4sf)(__m128)(b), (c)))

/// Compares each of the corresponding scalar values of two 128-bit
///    vectors of [4 x float], using the operation specified by the immediate
///    integer operand.
///
///    Each comparison returns 0x0 for false, 0xFFFFFFFF for true.
///    If either value in a comparison is NaN, comparisons that are ordered
///    return false, and comparisons that are unordered return true.
///
/// \headerfile <x86intrin.h>
///
/// \code
/// __m128 _mm_cmp_ss(__m128 a, __m128 b, const int c);
/// \endcode
///
/// This intrinsic corresponds to the <c> (V)CMPSS </c> instruction.
///
/// \param a
///    A 128-bit vector of [4 x float].
/// \param b
///    A 128-bit vector of [4 x float].
/// \param c
///    An immediate integer operand, with bits [4:0] specifying which comparison
///    operation to use: \n
///    0x00: Equal (ordered, non-signaling) \n
///    0x01: Less-than (ordered, signaling) \n
///    0x02: Less-than-or-equal (ordered, signaling) \n
///    0x03: Unordered (non-signaling) \n
///    0x04: Not-equal (unordered, non-signaling) \n
///    0x05: Not-less-than (unordered, signaling) \n
///    0x06: Not-less-than-or-equal (unordered, signaling) \n
///    0x07: Ordered (non-signaling) \n
/// \returns A 128-bit vector of [4 x float] containing the comparison results.
#define _mm_cmp_ss(a, b, c)                                                    \
  ((__m128)__builtin_ia32_cmpss((__v4sf)(__m128)(a), (__v4sf)(__m128)(b), (c)))

#define _MM_ALIGN16 __attribute__((aligned(16)))

#define _MM_SHUFFLE(z, y, x, w) (((z) << 6) | ((y) << 4) | ((x) << 2) | (w))

#define _MM_EXCEPT_INVALID    (0x0001U)
#define _MM_EXCEPT_DENORM     (0x0002U)
#define _MM_EXCEPT_DIV_ZERO   (0x0004U)
#define _MM_EXCEPT_OVERFLOW   (0x0008U)
#define _MM_EXCEPT_UNDERFLOW  (0x0010U)
#define _MM_EXCEPT_INEXACT    (0x0020U)
#define _MM_EXCEPT_MASK       (0x003fU)

#define _MM_MASK_INVALID      (0x0080U)
#define _MM_MASK_DENORM       (0x0100U)
#define _MM_MASK_DIV_ZERO     (0x0200U)
#define _MM_MASK_OVERFLOW     (0x0400U)
#define _MM_MASK_UNDERFLOW    (0x0800U)
#define _MM_MASK_INEXACT      (0x1000U)
#define _MM_MASK_MASK         (0x1f80U)

#define _MM_ROUND_NEAREST     (0x0000U)
#define _MM_ROUND_DOWN        (0x2000U)
#define _MM_ROUND_UP          (0x4000U)
#define _MM_ROUND_TOWARD_ZERO (0x6000U)
#define _MM_ROUND_MASK        (0x6000U)

#define _MM_FLUSH_ZERO_MASK   (0x8000U)
#define _MM_FLUSH_ZERO_ON     (0x8000U)
#define _MM_FLUSH_ZERO_OFF    (0x0000U)

#define _MM_GET_EXCEPTION_MASK() (_mm_getcsr() & _MM_MASK_MASK)
#define _MM_GET_EXCEPTION_STATE() (_mm_getcsr() & _MM_EXCEPT_MASK)
#define _MM_GET_FLUSH_ZERO_MODE() (_mm_getcsr() & _MM_FLUSH_ZERO_MASK)
#define _MM_GET_ROUNDING_MODE() (_mm_getcsr() & _MM_ROUND_MASK)

#define _MM_SET_EXCEPTION_MASK(x) (_mm_setcsr((_mm_getcsr() & ~_MM_MASK_MASK) | (x)))
#define _MM_SET_EXCEPTION_STATE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_EXCEPT_MASK) | (x)))
#define _MM_SET_FLUSH_ZERO_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_FLUSH_ZERO_MASK) | (x)))
#define _MM_SET_ROUNDING_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_ROUND_MASK) | (x)))

#define _MM_TRANSPOSE4_PS(row0, row1, row2, row3) \
do { \
  __m128 tmp3, tmp2, tmp1, tmp0; \
  tmp0 = _mm_unpacklo_ps((row0), (row1)); \
  tmp2 = _mm_unpacklo_ps((row2), (row3)); \
  tmp1 = _mm_unpackhi_ps((row0), (row1)); \
  tmp3 = _mm_unpackhi_ps((row2), (row3)); \
  (row0) = _mm_movelh_ps(tmp0, tmp2); \
  (row1) = _mm_movehl_ps(tmp2, tmp0); \
  (row2) = _mm_movelh_ps(tmp1, tmp3); \
  (row3) = _mm_movehl_ps(tmp3, tmp1); \
} while (0)

/* Aliases for compatibility. */
#define _m_pextrw _mm_extract_pi16
#define _m_pinsrw _mm_insert_pi16
#define _m_pmaxsw _mm_max_pi16
#define _m_pmaxub _mm_max_pu8
#define _m_pminsw _mm_min_pi16
#define _m_pminub _mm_min_pu8
#define _m_pmovmskb _mm_movemask_pi8
#define _m_pmulhuw _mm_mulhi_pu16
#define _m_pshufw _mm_shuffle_pi16
#define _m_maskmovq _mm_maskmove_si64
#define _m_pavgb _mm_avg_pu8
#define _m_pavgw _mm_avg_pu16
#define _m_psadbw _mm_sad_pu8
#define _m_ _mm_

#undef __trunc64
#undef __zext128
#undef __anyext128
#undef __zeroupper64
#undef __DEFAULT_FN_ATTRS
#undef __DEFAULT_FN_ATTRS_CONSTEXPR
#undef __DEFAULT_FN_ATTRS_SSE2
#undef __DEFAULT_FN_ATTRS_SSE2_CONSTEXPR

/* Ugly hack for backwards-compatibility (compatible with gcc) */
#if defined(__SSE2__) && !__building_module(_Builtin_intrinsics)
#include <emmintrin.h>
#endif

#endif /* __XMMINTRIN_H */

Youez - 2016 - github.com/yon3zu
LinuXploit