1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414
//! Streaming SIMD Extensions (SSE)
use crate::{
core_arch::{simd::*, x86::*},
intrinsics::simd::*,
intrinsics::sqrtf32,
mem, ptr,
};
#[cfg(test)]
use stdarch_test::assert_instr;
/// Adds the first component of `a` and `b`, the other components are copied
/// from `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_add_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(addss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_add_ss(a: __m128, b: __m128) -> __m128 {
simd_insert!(a, 0, _mm_cvtss_f32(a) + _mm_cvtss_f32(b))
}
/// Adds __m128 vectors.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_add_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(addps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_add_ps(a: __m128, b: __m128) -> __m128 {
simd_add(a, b)
}
/// Subtracts the first component of `b` from `a`, the other components are
/// copied from `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_sub_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(subss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_sub_ss(a: __m128, b: __m128) -> __m128 {
simd_insert!(a, 0, _mm_cvtss_f32(a) - _mm_cvtss_f32(b))
}
/// Subtracts __m128 vectors.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_sub_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(subps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_sub_ps(a: __m128, b: __m128) -> __m128 {
simd_sub(a, b)
}
/// Multiplies the first component of `a` and `b`, the other components are
/// copied from `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mul_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(mulss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_mul_ss(a: __m128, b: __m128) -> __m128 {
simd_insert!(a, 0, _mm_cvtss_f32(a) * _mm_cvtss_f32(b))
}
/// Multiplies __m128 vectors.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mul_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(mulps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_mul_ps(a: __m128, b: __m128) -> __m128 {
simd_mul(a, b)
}
/// Divides the first component of `b` by `a`, the other components are
/// copied from `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_div_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(divss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_div_ss(a: __m128, b: __m128) -> __m128 {
simd_insert!(a, 0, _mm_cvtss_f32(a) / _mm_cvtss_f32(b))
}
/// Divides __m128 vectors.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_div_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(divps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_div_ps(a: __m128, b: __m128) -> __m128 {
simd_div(a, b)
}
/// Returns the square root of the first single-precision (32-bit)
/// floating-point element in `a`, the other elements are unchanged.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_sqrt_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(sqrtss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_sqrt_ss(a: __m128) -> __m128 {
simd_insert!(a, 0, sqrtf32(_mm_cvtss_f32(a)))
}
/// Returns the square root of packed single-precision (32-bit) floating-point
/// elements in `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_sqrt_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(sqrtps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_sqrt_ps(a: __m128) -> __m128 {
simd_fsqrt(a)
}
/// Returns the approximate reciprocal of the first single-precision
/// (32-bit) floating-point element in `a`, the other elements are unchanged.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_rcp_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(rcpss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_rcp_ss(a: __m128) -> __m128 {
rcpss(a)
}
/// Returns the approximate reciprocal of packed single-precision (32-bit)
/// floating-point elements in `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_rcp_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(rcpps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_rcp_ps(a: __m128) -> __m128 {
rcpps(a)
}
/// Returns the approximate reciprocal square root of the first single-precision
/// (32-bit) floating-point element in `a`, the other elements are unchanged.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_rsqrt_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(rsqrtss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_rsqrt_ss(a: __m128) -> __m128 {
rsqrtss(a)
}
/// Returns the approximate reciprocal square root of packed single-precision
/// (32-bit) floating-point elements in `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_rsqrt_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(rsqrtps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_rsqrt_ps(a: __m128) -> __m128 {
rsqrtps(a)
}
/// Compares the first single-precision (32-bit) floating-point element of `a`
/// and `b`, and return the minimum value in the first element of the return
/// value, the other elements are copied from `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_min_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(minss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_min_ss(a: __m128, b: __m128) -> __m128 {
minss(a, b)
}
/// Compares packed single-precision (32-bit) floating-point elements in `a` and
/// `b`, and return the corresponding minimum values.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_min_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(minps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_min_ps(a: __m128, b: __m128) -> __m128 {
// See the `test_mm_min_ps` test why this can't be implemented using `simd_fmin`.
minps(a, b)
}
/// Compares the first single-precision (32-bit) floating-point element of `a`
/// and `b`, and return the maximum value in the first element of the return
/// value, the other elements are copied from `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_max_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(maxss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_max_ss(a: __m128, b: __m128) -> __m128 {
maxss(a, b)
}
/// Compares packed single-precision (32-bit) floating-point elements in `a` and
/// `b`, and return the corresponding maximum values.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_max_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(maxps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_max_ps(a: __m128, b: __m128) -> __m128 {
// See the `test_mm_min_ps` test why this can't be implemented using `simd_fmax`.
maxps(a, b)
}
/// Bitwise AND of packed single-precision (32-bit) floating-point elements.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_and_ps)
#[inline]
#[target_feature(enable = "sse")]
// i586 only seems to generate plain `and` instructions, so ignore it.
#[cfg_attr(
all(test, any(target_arch = "x86_64", target_feature = "sse2")),
assert_instr(andps)
)]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_and_ps(a: __m128, b: __m128) -> __m128 {
let a: __m128i = mem::transmute(a);
let b: __m128i = mem::transmute(b);
mem::transmute(simd_and(a, b))
}
/// Bitwise AND-NOT of packed single-precision (32-bit) floating-point
/// elements.
///
/// Computes `!a & b` for each bit in `a` and `b`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_andnot_ps)
#[inline]
#[target_feature(enable = "sse")]
// i586 only seems to generate plain `not` and `and` instructions, so ignore
// it.
#[cfg_attr(
all(test, any(target_arch = "x86_64", target_feature = "sse2")),
assert_instr(andnps)
)]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_andnot_ps(a: __m128, b: __m128) -> __m128 {
let a: __m128i = mem::transmute(a);
let b: __m128i = mem::transmute(b);
let mask: __m128i = mem::transmute(i32x4::splat(-1));
mem::transmute(simd_and(simd_xor(mask, a), b))
}
/// Bitwise OR of packed single-precision (32-bit) floating-point elements.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_or_ps)
#[inline]
#[target_feature(enable = "sse")]
// i586 only seems to generate plain `or` instructions, so we ignore it.
#[cfg_attr(
all(test, any(target_arch = "x86_64", target_feature = "sse2")),
assert_instr(orps)
)]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_or_ps(a: __m128, b: __m128) -> __m128 {
let a: __m128i = mem::transmute(a);
let b: __m128i = mem::transmute(b);
mem::transmute(simd_or(a, b))
}
/// Bitwise exclusive OR of packed single-precision (32-bit) floating-point
/// elements.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_xor_ps)
#[inline]
#[target_feature(enable = "sse")]
// i586 only seems to generate plain `xor` instructions, so we ignore it.
#[cfg_attr(
all(test, any(target_arch = "x86_64", target_feature = "sse2")),
assert_instr(xorps)
)]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_xor_ps(a: __m128, b: __m128) -> __m128 {
let a: __m128i = mem::transmute(a);
let b: __m128i = mem::transmute(b);
mem::transmute(simd_xor(a, b))
}
/// Compares the lowest `f32` of both inputs for equality. The lowest 32 bits of
/// the result will be `0xffffffff` if the two inputs are equal, or `0`
/// otherwise. The upper 96 bits of the result are the upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpeq_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpeqss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpeq_ss(a: __m128, b: __m128) -> __m128 {
cmpss(a, b, 0)
}
/// Compares the lowest `f32` of both inputs for less than. The lowest 32 bits
/// of the result will be `0xffffffff` if `a.extract(0)` is less than
/// `b.extract(0)`, or `0` otherwise. The upper 96 bits of the result are the
/// upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmplt_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpltss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmplt_ss(a: __m128, b: __m128) -> __m128 {
cmpss(a, b, 1)
}
/// Compares the lowest `f32` of both inputs for less than or equal. The lowest
/// 32 bits of the result will be `0xffffffff` if `a.extract(0)` is less than
/// or equal `b.extract(0)`, or `0` otherwise. The upper 96 bits of the result
/// are the upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmple_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpless))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmple_ss(a: __m128, b: __m128) -> __m128 {
cmpss(a, b, 2)
}
/// Compares the lowest `f32` of both inputs for greater than. The lowest 32
/// bits of the result will be `0xffffffff` if `a.extract(0)` is greater
/// than `b.extract(0)`, or `0` otherwise. The upper 96 bits of the result
/// are the upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpgt_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpltss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpgt_ss(a: __m128, b: __m128) -> __m128 {
simd_shuffle!(a, cmpss(b, a, 1), [4, 1, 2, 3])
}
/// Compares the lowest `f32` of both inputs for greater than or equal. The
/// lowest 32 bits of the result will be `0xffffffff` if `a.extract(0)` is
/// greater than or equal `b.extract(0)`, or `0` otherwise. The upper 96 bits
/// of the result are the upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpge_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpless))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpge_ss(a: __m128, b: __m128) -> __m128 {
simd_shuffle!(a, cmpss(b, a, 2), [4, 1, 2, 3])
}
/// Compares the lowest `f32` of both inputs for inequality. The lowest 32 bits
/// of the result will be `0xffffffff` if `a.extract(0)` is not equal to
/// `b.extract(0)`, or `0` otherwise. The upper 96 bits of the result are the
/// upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpneq_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpneqss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpneq_ss(a: __m128, b: __m128) -> __m128 {
cmpss(a, b, 4)
}
/// Compares the lowest `f32` of both inputs for not-less-than. The lowest 32
/// bits of the result will be `0xffffffff` if `a.extract(0)` is not less than
/// `b.extract(0)`, or `0` otherwise. The upper 96 bits of the result are the
/// upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpnlt_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpnltss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpnlt_ss(a: __m128, b: __m128) -> __m128 {
cmpss(a, b, 5)
}
/// Compares the lowest `f32` of both inputs for not-less-than-or-equal. The
/// lowest 32 bits of the result will be `0xffffffff` if `a.extract(0)` is not
/// less than or equal to `b.extract(0)`, or `0` otherwise. The upper 96 bits
/// of the result are the upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpnle_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpnless))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpnle_ss(a: __m128, b: __m128) -> __m128 {
cmpss(a, b, 6)
}
/// Compares the lowest `f32` of both inputs for not-greater-than. The lowest 32
/// bits of the result will be `0xffffffff` if `a.extract(0)` is not greater
/// than `b.extract(0)`, or `0` otherwise. The upper 96 bits of the result are
/// the upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpngt_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpnltss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpngt_ss(a: __m128, b: __m128) -> __m128 {
simd_shuffle!(a, cmpss(b, a, 5), [4, 1, 2, 3])
}
/// Compares the lowest `f32` of both inputs for not-greater-than-or-equal. The
/// lowest 32 bits of the result will be `0xffffffff` if `a.extract(0)` is not
/// greater than or equal to `b.extract(0)`, or `0` otherwise. The upper 96
/// bits of the result are the upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpnge_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpnless))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpnge_ss(a: __m128, b: __m128) -> __m128 {
simd_shuffle!(a, cmpss(b, a, 6), [4, 1, 2, 3])
}
/// Checks if the lowest `f32` of both inputs are ordered. The lowest 32 bits of
/// the result will be `0xffffffff` if neither of `a.extract(0)` or
/// `b.extract(0)` is a NaN, or `0` otherwise. The upper 96 bits of the result
/// are the upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpord_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpordss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpord_ss(a: __m128, b: __m128) -> __m128 {
cmpss(a, b, 7)
}
/// Checks if the lowest `f32` of both inputs are unordered. The lowest 32 bits
/// of the result will be `0xffffffff` if any of `a.extract(0)` or
/// `b.extract(0)` is a NaN, or `0` otherwise. The upper 96 bits of the result
/// are the upper 96 bits of `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpunord_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpunordss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpunord_ss(a: __m128, b: __m128) -> __m128 {
cmpss(a, b, 3)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// The result in the output vector will be `0xffffffff` if the input elements
/// were equal, or `0` otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpeq_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpeqps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpeq_ps(a: __m128, b: __m128) -> __m128 {
cmpps(a, b, 0)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// The result in the output vector will be `0xffffffff` if the input element
/// in `a` is less than the corresponding element in `b`, or `0` otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmplt_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpltps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmplt_ps(a: __m128, b: __m128) -> __m128 {
cmpps(a, b, 1)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// The result in the output vector will be `0xffffffff` if the input element
/// in `a` is less than or equal to the corresponding element in `b`, or `0`
/// otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmple_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpleps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmple_ps(a: __m128, b: __m128) -> __m128 {
cmpps(a, b, 2)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// The result in the output vector will be `0xffffffff` if the input element
/// in `a` is greater than the corresponding element in `b`, or `0` otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpgt_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpltps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpgt_ps(a: __m128, b: __m128) -> __m128 {
cmpps(b, a, 1)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// The result in the output vector will be `0xffffffff` if the input element
/// in `a` is greater than or equal to the corresponding element in `b`, or `0`
/// otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpge_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpleps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpge_ps(a: __m128, b: __m128) -> __m128 {
cmpps(b, a, 2)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// The result in the output vector will be `0xffffffff` if the input elements
/// are **not** equal, or `0` otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpneq_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpneqps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpneq_ps(a: __m128, b: __m128) -> __m128 {
cmpps(a, b, 4)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// The result in the output vector will be `0xffffffff` if the input element
/// in `a` is **not** less than the corresponding element in `b`, or `0`
/// otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpnlt_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpnltps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpnlt_ps(a: __m128, b: __m128) -> __m128 {
cmpps(a, b, 5)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// The result in the output vector will be `0xffffffff` if the input element
/// in `a` is **not** less than or equal to the corresponding element in `b`, or
/// `0` otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpnle_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpnleps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpnle_ps(a: __m128, b: __m128) -> __m128 {
cmpps(a, b, 6)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// The result in the output vector will be `0xffffffff` if the input element
/// in `a` is **not** greater than the corresponding element in `b`, or `0`
/// otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpngt_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpnltps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpngt_ps(a: __m128, b: __m128) -> __m128 {
cmpps(b, a, 5)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// The result in the output vector will be `0xffffffff` if the input element
/// in `a` is **not** greater than or equal to the corresponding element in `b`,
/// or `0` otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpnge_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpnleps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpnge_ps(a: __m128, b: __m128) -> __m128 {
cmpps(b, a, 6)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// Returns four floats that have one of two possible bit patterns. The element
/// in the output vector will be `0xffffffff` if the input elements in `a` and
/// `b` are ordered (i.e., neither of them is a NaN), or 0 otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpord_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpordps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpord_ps(a: __m128, b: __m128) -> __m128 {
cmpps(b, a, 7)
}
/// Compares each of the four floats in `a` to the corresponding element in `b`.
/// Returns four floats that have one of two possible bit patterns. The element
/// in the output vector will be `0xffffffff` if the input elements in `a` and
/// `b` are unordered (i.e., at least on of them is a NaN), or 0 otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cmpunord_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cmpunordps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cmpunord_ps(a: __m128, b: __m128) -> __m128 {
cmpps(b, a, 3)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if they are equal, or `0` otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_comieq_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(comiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_comieq_ss(a: __m128, b: __m128) -> i32 {
comieq_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if the value from `a` is less than the one from `b`, or `0` otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_comilt_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(comiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_comilt_ss(a: __m128, b: __m128) -> i32 {
comilt_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if the value from `a` is less than or equal to the one from `b`, or `0`
/// otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_comile_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(comiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_comile_ss(a: __m128, b: __m128) -> i32 {
comile_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if the value from `a` is greater than the one from `b`, or `0`
/// otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_comigt_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(comiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_comigt_ss(a: __m128, b: __m128) -> i32 {
comigt_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if the value from `a` is greater than or equal to the one from `b`, or
/// `0` otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_comige_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(comiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_comige_ss(a: __m128, b: __m128) -> i32 {
comige_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if they are **not** equal, or `0` otherwise.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_comineq_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(comiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_comineq_ss(a: __m128, b: __m128) -> i32 {
comineq_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if they are equal, or `0` otherwise. This instruction will not signal
/// an exception if either argument is a quiet NaN.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_ucomieq_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(ucomiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_ucomieq_ss(a: __m128, b: __m128) -> i32 {
ucomieq_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if the value from `a` is less than the one from `b`, or `0` otherwise.
/// This instruction will not signal an exception if either argument is a quiet
/// NaN.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_ucomilt_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(ucomiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_ucomilt_ss(a: __m128, b: __m128) -> i32 {
ucomilt_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if the value from `a` is less than or equal to the one from `b`, or `0`
/// otherwise. This instruction will not signal an exception if either argument
/// is a quiet NaN.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_ucomile_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(ucomiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_ucomile_ss(a: __m128, b: __m128) -> i32 {
ucomile_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if the value from `a` is greater than the one from `b`, or `0`
/// otherwise. This instruction will not signal an exception if either argument
/// is a quiet NaN.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_ucomigt_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(ucomiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_ucomigt_ss(a: __m128, b: __m128) -> i32 {
ucomigt_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if the value from `a` is greater than or equal to the one from `b`, or
/// `0` otherwise. This instruction will not signal an exception if either
/// argument is a quiet NaN.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_ucomige_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(ucomiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_ucomige_ss(a: __m128, b: __m128) -> i32 {
ucomige_ss(a, b)
}
/// Compares two 32-bit floats from the low-order bits of `a` and `b`. Returns
/// `1` if they are **not** equal, or `0` otherwise. This instruction will not
/// signal an exception if either argument is a quiet NaN.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_ucomineq_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(ucomiss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_ucomineq_ss(a: __m128, b: __m128) -> i32 {
ucomineq_ss(a, b)
}
/// Converts the lowest 32 bit float in the input vector to a 32 bit integer.
///
/// The result is rounded according to the current rounding mode. If the result
/// cannot be represented as a 32 bit integer the result will be `0x8000_0000`
/// (`i32::MIN`).
///
/// This corresponds to the `CVTSS2SI` instruction (with 32 bit output).
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtss_si32)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cvtss2si))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cvtss_si32(a: __m128) -> i32 {
cvtss2si(a)
}
/// Alias for [`_mm_cvtss_si32`](fn._mm_cvtss_si32.html).
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_ss2si)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cvtss2si))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cvt_ss2si(a: __m128) -> i32 {
_mm_cvtss_si32(a)
}
/// Converts the lowest 32 bit float in the input vector to a 32 bit integer
/// with
/// truncation.
///
/// The result is rounded always using truncation (round towards zero). If the
/// result cannot be represented as a 32 bit integer the result will be
/// `0x8000_0000` (`i32::MIN`).
///
/// This corresponds to the `CVTTSS2SI` instruction (with 32 bit output).
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvttss_si32)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cvttss2si))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cvttss_si32(a: __m128) -> i32 {
cvttss2si(a)
}
/// Alias for [`_mm_cvttss_si32`](fn._mm_cvttss_si32.html).
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtt_ss2si)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cvttss2si))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cvtt_ss2si(a: __m128) -> i32 {
_mm_cvttss_si32(a)
}
/// Extracts the lowest 32 bit float from the input vector.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtss_f32)
#[inline]
#[target_feature(enable = "sse")]
// No point in using assert_instrs. In Unix x86_64 calling convention this is a
// no-op, and on msvc it's just a `mov`.
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cvtss_f32(a: __m128) -> f32 {
simd_extract!(a, 0)
}
/// Converts a 32 bit integer to a 32 bit float. The result vector is the input
/// vector `a` with the lowest 32 bit float replaced by the converted integer.
///
/// This intrinsic corresponds to the `CVTSI2SS` instruction (with 32 bit
/// input).
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtsi32_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cvtsi2ss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cvtsi32_ss(a: __m128, b: i32) -> __m128 {
cvtsi2ss(a, b)
}
/// Alias for [`_mm_cvtsi32_ss`](fn._mm_cvtsi32_ss.html).
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvt_si2ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(cvtsi2ss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_cvt_si2ss(a: __m128, b: i32) -> __m128 {
_mm_cvtsi32_ss(a, b)
}
/// Construct a `__m128` with the lowest element set to `a` and the rest set to
/// zero.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_set_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_set_ss(a: f32) -> __m128 {
__m128([a, 0.0, 0.0, 0.0])
}
/// Construct a `__m128` with all element set to `a`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_set1_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(shufps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_set1_ps(a: f32) -> __m128 {
__m128([a, a, a, a])
}
/// Alias for [`_mm_set1_ps`](fn._mm_set1_ps.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_set_ps1)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(shufps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_set_ps1(a: f32) -> __m128 {
_mm_set1_ps(a)
}
/// Construct a `__m128` from four floating point values highest to lowest.
///
/// Note that `a` will be the highest 32 bits of the result, and `d` the
/// lowest. This matches the standard way of writing bit patterns on x86:
///
/// ```text
/// bit 127 .. 96 95 .. 64 63 .. 32 31 .. 0
/// +---------+---------+---------+---------+
/// | a | b | c | d | result
/// +---------+---------+---------+---------+
/// ```
///
/// Alternatively:
///
/// ```text
/// let v = _mm_set_ps(d, c, b, a);
/// ```
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_set_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(unpcklps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_set_ps(a: f32, b: f32, c: f32, d: f32) -> __m128 {
__m128([d, c, b, a])
}
/// Construct a `__m128` from four floating point values lowest to highest.
///
/// This matches the memory order of `__m128`, i.e., `a` will be the lowest 32
/// bits of the result, and `d` the highest.
///
/// ```text
/// assert_eq!(__m128::new(a, b, c, d), _mm_setr_ps(a, b, c, d));
/// ```
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_setr_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(
all(test, any(target_env = "msvc", target_arch = "x86_64")),
assert_instr(unpcklps)
)]
// On a 32-bit architecture on non-msvc it just copies the operands from the stack.
#[cfg_attr(
all(test, all(not(target_env = "msvc"), target_arch = "x86")),
assert_instr(movaps)
)]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_setr_ps(a: f32, b: f32, c: f32, d: f32) -> __m128 {
__m128([a, b, c, d])
}
/// Construct a `__m128` with all elements initialized to zero.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_setzero_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(xorps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_setzero_ps() -> __m128 {
__m128([0.0, 0.0, 0.0, 0.0])
}
/// A utility function for creating masks to use with Intel shuffle and
/// permute intrinsics.
#[inline]
#[allow(non_snake_case)]
#[unstable(feature = "stdarch_x86_mm_shuffle", issue = "111147")]
pub const fn _MM_SHUFFLE(z: u32, y: u32, x: u32, w: u32) -> i32 {
((z << 6) | (y << 4) | (x << 2) | w) as i32
}
/// Shuffles packed single-precision (32-bit) floating-point elements in `a` and
/// `b` using `MASK`.
///
/// The lower half of result takes values from `a` and the higher half from
/// `b`. Mask is split to 2 control bits each to index the element from inputs.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_shuffle_ps)
///
/// Note that there appears to be a mistake within Intel's Intrinsics Guide.
/// `_mm_shuffle_ps` is supposed to take an `i32` instead of a `u32`
/// as is the case for [other shuffle intrinsics](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_shuffle_).
/// Performing an implicit type conversion between an unsigned integer and a signed integer
/// does not cause a problem in C, however Rust's commitment to strong typing does not allow this.
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(shufps, MASK = 3))]
#[rustc_legacy_const_generics(2)]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_shuffle_ps<const MASK: i32>(a: __m128, b: __m128) -> __m128 {
static_assert_uimm_bits!(MASK, 8);
simd_shuffle!(
a,
b,
[
MASK as u32 & 0b11,
(MASK as u32 >> 2) & 0b11,
((MASK as u32 >> 4) & 0b11) + 4,
((MASK as u32 >> 6) & 0b11) + 4,
],
)
}
/// Unpacks and interleave single-precision (32-bit) floating-point elements
/// from the higher half of `a` and `b`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_unpackhi_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(unpckhps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_unpackhi_ps(a: __m128, b: __m128) -> __m128 {
simd_shuffle!(a, b, [2, 6, 3, 7])
}
/// Unpacks and interleave single-precision (32-bit) floating-point elements
/// from the lower half of `a` and `b`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_unpacklo_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(unpcklps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_unpacklo_ps(a: __m128, b: __m128) -> __m128 {
simd_shuffle!(a, b, [0, 4, 1, 5])
}
/// Combine higher half of `a` and `b`. The higher half of `b` occupies the
/// lower half of result.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movehl_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(all(test, not(target_env = "msvc")), assert_instr(movhlps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_movehl_ps(a: __m128, b: __m128) -> __m128 {
// TODO; figure why this is a different instruction on msvc?
simd_shuffle!(a, b, [6, 7, 2, 3])
}
/// Combine lower half of `a` and `b`. The lower half of `b` occupies the
/// higher half of result.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movelh_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(all(test, not(target_env = "msvc")), assert_instr(movlhps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_movelh_ps(a: __m128, b: __m128) -> __m128 {
simd_shuffle!(a, b, [0, 1, 4, 5])
}
/// Returns a mask of the most significant bit of each element in `a`.
///
/// The mask is stored in the 4 least significant bits of the return value.
/// All other bits are set to `0`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_movemask_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movmskps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_movemask_ps(a: __m128) -> i32 {
// Propagate the highest bit to the rest, because simd_bitmask
// requires all-1 or all-0.
let mask: i32x4 = simd_lt(transmute(a), i32x4::splat(0));
simd_bitmask::<i32x4, u8>(mask).into()
}
/// Construct a `__m128` with the lowest element read from `p` and the other
/// elements set to zero.
///
/// This corresponds to instructions `VMOVSS` / `MOVSS`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_load_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_load_ss(p: *const f32) -> __m128 {
__m128([*p, 0.0, 0.0, 0.0])
}
/// Construct a `__m128` by duplicating the value read from `p` into all
/// elements.
///
/// This corresponds to instructions `VMOVSS` / `MOVSS` followed by some
/// shuffling.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_load1_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_load1_ps(p: *const f32) -> __m128 {
let a = *p;
__m128([a, a, a, a])
}
/// Alias for [`_mm_load1_ps`](fn._mm_load1_ps.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_load_ps1)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_load_ps1(p: *const f32) -> __m128 {
_mm_load1_ps(p)
}
/// Loads four `f32` values from *aligned* memory into a `__m128`. If the
/// pointer is not aligned to a 128-bit boundary (16 bytes) a general
/// protection fault will be triggered (fatal program crash).
///
/// Use [`_mm_loadu_ps`](fn._mm_loadu_ps.html) for potentially unaligned
/// memory.
///
/// This corresponds to instructions `VMOVAPS` / `MOVAPS`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_load_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movaps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[allow(clippy::cast_ptr_alignment)]
pub unsafe fn _mm_load_ps(p: *const f32) -> __m128 {
*(p as *const __m128)
}
/// Loads four `f32` values from memory into a `__m128`. There are no
/// restrictions
/// on memory alignment. For aligned memory
/// [`_mm_load_ps`](fn._mm_load_ps.html)
/// may be faster.
///
/// This corresponds to instructions `VMOVUPS` / `MOVUPS`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_loadu_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movups))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_loadu_ps(p: *const f32) -> __m128 {
// Note: Using `*p` would require `f32` alignment, but `movups` has no
// alignment restrictions.
let mut dst = _mm_undefined_ps();
ptr::copy_nonoverlapping(
p as *const u8,
ptr::addr_of_mut!(dst) as *mut u8,
mem::size_of::<__m128>(),
);
dst
}
/// Loads four `f32` values from aligned memory into a `__m128` in reverse
/// order.
///
/// If the pointer is not aligned to a 128-bit boundary (16 bytes) a general
/// protection fault will be triggered (fatal program crash).
///
/// Functionally equivalent to the following code sequence (assuming `p`
/// satisfies the alignment restrictions):
///
/// ```text
/// let a0 = *p;
/// let a1 = *p.add(1);
/// let a2 = *p.add(2);
/// let a3 = *p.add(3);
/// __m128::new(a3, a2, a1, a0)
/// ```
///
/// This corresponds to instructions `VMOVAPS` / `MOVAPS` followed by some
/// shuffling.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_loadr_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movaps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_loadr_ps(p: *const f32) -> __m128 {
let a = _mm_load_ps(p);
simd_shuffle!(a, a, [3, 2, 1, 0])
}
/// Stores the lowest 32 bit float of `a` into memory.
///
/// This intrinsic corresponds to the `MOVSS` instruction.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_store_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_store_ss(p: *mut f32, a: __m128) {
*p = simd_extract!(a, 0);
}
/// Stores the lowest 32 bit float of `a` repeated four times into *aligned*
/// memory.
///
/// If the pointer is not aligned to a 128-bit boundary (16 bytes) a general
/// protection fault will be triggered (fatal program crash).
///
/// Functionally equivalent to the following code sequence (assuming `p`
/// satisfies the alignment restrictions):
///
/// ```text
/// let x = a.extract(0);
/// *p = x;
/// *p.add(1) = x;
/// *p.add(2) = x;
/// *p.add(3) = x;
/// ```
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_store1_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movaps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[allow(clippy::cast_ptr_alignment)]
pub unsafe fn _mm_store1_ps(p: *mut f32, a: __m128) {
let b: __m128 = simd_shuffle!(a, a, [0, 0, 0, 0]);
*(p as *mut __m128) = b;
}
/// Alias for [`_mm_store1_ps`](fn._mm_store1_ps.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_store_ps1)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movaps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_store_ps1(p: *mut f32, a: __m128) {
_mm_store1_ps(p, a);
}
/// Stores four 32-bit floats into *aligned* memory.
///
/// If the pointer is not aligned to a 128-bit boundary (16 bytes) a general
/// protection fault will be triggered (fatal program crash).
///
/// Use [`_mm_storeu_ps`](fn._mm_storeu_ps.html) for potentially unaligned
/// memory.
///
/// This corresponds to instructions `VMOVAPS` / `MOVAPS`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_store_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movaps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[allow(clippy::cast_ptr_alignment)]
pub unsafe fn _mm_store_ps(p: *mut f32, a: __m128) {
*(p as *mut __m128) = a;
}
/// Stores four 32-bit floats into memory. There are no restrictions on memory
/// alignment. For aligned memory [`_mm_store_ps`](fn._mm_store_ps.html) may be
/// faster.
///
/// This corresponds to instructions `VMOVUPS` / `MOVUPS`.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_storeu_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movups))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_storeu_ps(p: *mut f32, a: __m128) {
ptr::copy_nonoverlapping(
ptr::addr_of!(a) as *const u8,
p as *mut u8,
mem::size_of::<__m128>(),
);
}
/// Stores four 32-bit floats into *aligned* memory in reverse order.
///
/// If the pointer is not aligned to a 128-bit boundary (16 bytes) a general
/// protection fault will be triggered (fatal program crash).
///
/// Functionally equivalent to the following code sequence (assuming `p`
/// satisfies the alignment restrictions):
///
/// ```text
/// *p = a.extract(3);
/// *p.add(1) = a.extract(2);
/// *p.add(2) = a.extract(1);
/// *p.add(3) = a.extract(0);
/// ```
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_storer_ps)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movaps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[allow(clippy::cast_ptr_alignment)]
pub unsafe fn _mm_storer_ps(p: *mut f32, a: __m128) {
let b: __m128 = simd_shuffle!(a, a, [3, 2, 1, 0]);
*(p as *mut __m128) = b;
}
/// Returns a `__m128` with the first component from `b` and the remaining
/// components from `a`.
///
/// In other words for any `a` and `b`:
/// ```text
/// _mm_move_ss(a, b) == a.replace(0, b.extract(0))
/// ```
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_move_ss)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movss))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_move_ss(a: __m128, b: __m128) -> __m128 {
simd_shuffle!(a, b, [4, 1, 2, 3])
}
/// Performs a serializing operation on all non-temporal ("streaming") store instructions that
/// were issued by the current thread prior to this instruction.
///
/// Guarantees that every non-temporal store instruction that precedes this fence, in program order, is
/// ordered before any load or store instruction which follows the fence in
/// synchronization order.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_sfence)
/// (but note that Intel is only documenting the hardware-level concerns related to this
/// instruction; the Intel documentation does not take into account the extra concerns that arise
/// because the Rust memory model is different from the x86 memory model.)
///
/// # Safety of non-temporal stores
///
/// After using any non-temporal store intrinsic, but before any other access to the memory that the
/// intrinsic mutates, a call to `_mm_sfence` must be performed on the thread that used the
/// intrinsic.
///
/// Non-temporal stores behave very different from regular stores. For the purpose of the Rust
/// memory model, these stores are happening asynchronously in a background thread. This means a
/// non-temporal store can cause data races with other accesses, even other accesses on the same
/// thread. It also means that cross-thread synchronization does not work as expected: let's say the
/// intrinsic is called on thread T1, and T1 performs synchronization with some other thread T2. The
/// non-temporal store acts as if it happened not in T1 but in a different thread T3, and T2 has not
/// synchronized with T3! Calling `_mm_sfence` makes the current thread wait for and synchronize
/// with all the non-temporal stores previously started on this thread, which means in particular
/// that subsequent synchronization with other threads will then work as intended again.
///
/// The general pattern to use non-temporal stores correctly is to call `_mm_sfence` before your
/// code jumps back to code outside your library. This ensures all stores inside your function
/// are synchronized-before the return, and thus transitively synchronized-before everything
/// the caller does after your function returns.
//
// The following is not a doc comment since it's not clear whether we want to put this into the
// docs, but it should be written out somewhere.
//
// Formally, we consider non-temporal stores and sfences to be opaque blobs that the compiler cannot
// inspect, and that behave like the following functions. This explains where the docs above come
// from.
// ```
// #[thread_local]
// static mut PENDING_NONTEMP_WRITES = AtomicUsize::new(0);
//
// pub unsafe fn nontemporal_store<T>(ptr: *mut T, val: T) {
// PENDING_NONTEMP_WRITES.fetch_add(1, Relaxed);
// // Spawn a thread that will eventually do our write.
// // We need to fetch a pointer to this thread's pending-write
// // counter, so that we can access it from the background thread.
// let pending_writes = addr_of!(PENDING_NONTEMP_WRITES);
// // If this was actual Rust code we'd have to do some extra work
// // because `ptr`, `val`, `pending_writes` are all `!Send`. We skip that here.
// std::thread::spawn(move || {
// // Do the write in the background thread.
// ptr.write(val);
// // Register the write as done. Crucially, this is `Release`, so it
// // syncs-with the `Acquire in `sfence`.
// (&*pending_writes).fetch_sub(1, Release);
// });
// }
//
// pub fn sfence() {
// unsafe {
// // Wait until there are no more pending writes.
// while PENDING_NONTEMP_WRITES.load(Acquire) > 0 {}
// }
// }
// ```
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(sfence))]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_sfence() {
sfence()
}
/// Gets the unsigned 32-bit value of the MXCSR control and status register.
///
/// Note that Rust makes no guarantees whatsoever about the contents of this register: Rust
/// floating-point operations may or may not result in this register getting updated with exception
/// state, and the register can change between two invocations of this function even when no
/// floating-point operations appear in the source code (since floating-point operations appearing
/// earlier or later can be reordered).
///
/// If you need to perform some floating-point operations and check whether they raised an
/// exception, use an inline assembly block for the entire sequence of operations.
///
/// For more info see [`_mm_setcsr`](fn._mm_setcsr.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_getcsr)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(stmxcsr))]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[deprecated(
since = "1.75.0",
note = "see `_mm_getcsr` documentation - use inline assembly instead"
)]
pub unsafe fn _mm_getcsr() -> u32 {
let mut result = 0_i32;
stmxcsr(ptr::addr_of_mut!(result) as *mut i8);
result as u32
}
/// Sets the MXCSR register with the 32-bit unsigned integer value.
///
/// This register controls how SIMD instructions handle floating point
/// operations. Modifying this register only affects the current thread.
///
/// It contains several groups of flags:
///
/// * *Exception flags* report which exceptions occurred since last they were
/// reset.
///
/// * *Masking flags* can be used to mask (ignore) certain exceptions. By
/// default
/// these flags are all set to 1, so all exceptions are masked. When an
/// an exception is masked, the processor simply sets the exception flag and
/// continues the operation. If the exception is unmasked, the flag is also set
/// but additionally an exception handler is invoked.
///
/// * *Rounding mode flags* control the rounding mode of floating point
/// instructions.
///
/// * The *denormals-are-zero mode flag* turns all numbers which would be
/// denormalized (exponent bits are all zeros) into zeros.
///
/// Note that modifying the masking flags, rounding mode, or denormals-are-zero mode flags leads to
/// **immediate Undefined Behavior**: Rust assumes that these are always in their default state and
/// will optimize accordingly. This even applies when the register is altered and later reset to its
/// original value without any floating-point operations appearing in the source code between those
/// operations (since floating-point operations appearing earlier or later can be reordered).
///
/// If you need to perform some floating-point operations under a different masking flags, rounding
/// mode, or denormals-are-zero mode, use an inline assembly block and make sure to restore the
/// original MXCSR register state before the end of the block.
///
/// ## Exception Flags
///
/// * `_MM_EXCEPT_INVALID`: An invalid operation was performed (e.g., dividing
/// Infinity by Infinity).
///
/// * `_MM_EXCEPT_DENORM`: An operation attempted to operate on a denormalized
/// number. Mainly this can cause loss of precision.
///
/// * `_MM_EXCEPT_DIV_ZERO`: Division by zero occurred.
///
/// * `_MM_EXCEPT_OVERFLOW`: A numeric overflow exception occurred, i.e., a
/// result was too large to be represented (e.g., an `f32` with absolute
/// value
/// greater than `2^128`).
///
/// * `_MM_EXCEPT_UNDERFLOW`: A numeric underflow exception occurred, i.e., a
/// result was too small to be represented in a normalized way (e.g., an
/// `f32`
/// with absulte value smaller than `2^-126`.)
///
/// * `_MM_EXCEPT_INEXACT`: An inexact-result exception occurred (a.k.a.
/// precision exception). This means some precision was lost due to rounding.
/// For example, the fraction `1/3` cannot be represented accurately in a
/// 32 or 64 bit float and computing it would cause this exception to be
/// raised. Precision exceptions are very common, so they are usually masked.
///
/// Exception flags can be read and set using the convenience functions
/// `_MM_GET_EXCEPTION_STATE` and `_MM_SET_EXCEPTION_STATE`. For example, to
/// check if an operation caused some overflow:
///
/// ```rust,ignore
/// _MM_SET_EXCEPTION_STATE(0); // clear all exception flags
/// // perform calculations
/// if _MM_GET_EXCEPTION_STATE() & _MM_EXCEPT_OVERFLOW != 0 {
/// // handle overflow
/// }
/// ```
///
/// ## Masking Flags
///
/// There is one masking flag for each exception flag: `_MM_MASK_INVALID`,
/// `_MM_MASK_DENORM`, `_MM_MASK_DIV_ZERO`, `_MM_MASK_OVERFLOW`,
/// `_MM_MASK_UNDERFLOW`, `_MM_MASK_INEXACT`.
///
/// A single masking bit can be set via
///
/// ```rust,ignore
/// _MM_SET_EXCEPTION_MASK(_MM_MASK_UNDERFLOW);
/// ```
///
/// However, since mask bits are by default all set to 1, it is more common to
/// want to *disable* certain bits. For example, to unmask the underflow
/// exception, use:
///
/// ```rust,ignore
/// _mm_setcsr(_mm_getcsr() & !_MM_MASK_UNDERFLOW); // unmask underflow
/// exception
/// ```
///
/// Warning: an unmasked exception will cause an exception handler to be
/// called.
/// The standard handler will simply terminate the process. So, in this case
/// any underflow exception would terminate the current process with something
/// like `signal: 8, SIGFPE: erroneous arithmetic operation`.
///
/// ## Rounding Mode
///
/// The rounding mode is describe using two bits. It can be read and set using
/// the convenience wrappers `_MM_GET_ROUNDING_MODE()` and
/// `_MM_SET_ROUNDING_MODE(mode)`.
///
/// The rounding modes are:
///
/// * `_MM_ROUND_NEAREST`: (default) Round to closest to the infinite precision
/// value. If two values are equally close, round to even (i.e., least
/// significant bit will be zero).
///
/// * `_MM_ROUND_DOWN`: Round toward negative Infinity.
///
/// * `_MM_ROUND_UP`: Round toward positive Infinity.
///
/// * `_MM_ROUND_TOWARD_ZERO`: Round towards zero (truncate).
///
/// Example:
///
/// ```rust,ignore
/// _MM_SET_ROUNDING_MODE(_MM_ROUND_DOWN)
/// ```
///
/// ## Denormals-are-zero/Flush-to-zero Mode
///
/// If this bit is set, values that would be denormalized will be set to zero
/// instead. This is turned off by default.
///
/// You can read and enable/disable this mode via the helper functions
/// `_MM_GET_FLUSH_ZERO_MODE()` and `_MM_SET_FLUSH_ZERO_MODE()`:
///
/// ```rust,ignore
/// _MM_SET_FLUSH_ZERO_MODE(_MM_FLUSH_ZERO_OFF); // turn off (default)
/// _MM_SET_FLUSH_ZERO_MODE(_MM_FLUSH_ZERO_ON); // turn on
/// ```
///
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_setcsr)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(ldmxcsr))]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[deprecated(
since = "1.75.0",
note = "see `_mm_setcsr` documentation - use inline assembly instead"
)]
pub unsafe fn _mm_setcsr(val: u32) {
ldmxcsr(ptr::addr_of!(val) as *const i8);
}
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_EXCEPT_INVALID: u32 = 0x0001;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_EXCEPT_DENORM: u32 = 0x0002;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_EXCEPT_DIV_ZERO: u32 = 0x0004;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_EXCEPT_OVERFLOW: u32 = 0x0008;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_EXCEPT_UNDERFLOW: u32 = 0x0010;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_EXCEPT_INEXACT: u32 = 0x0020;
/// See [`_MM_GET_EXCEPTION_STATE`](fn._MM_GET_EXCEPTION_STATE.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_EXCEPT_MASK: u32 = 0x003f;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_MASK_INVALID: u32 = 0x0080;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_MASK_DENORM: u32 = 0x0100;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_MASK_DIV_ZERO: u32 = 0x0200;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_MASK_OVERFLOW: u32 = 0x0400;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_MASK_UNDERFLOW: u32 = 0x0800;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_MASK_INEXACT: u32 = 0x1000;
/// See [`_MM_GET_EXCEPTION_MASK`](fn._MM_GET_EXCEPTION_MASK.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_MASK_MASK: u32 = 0x1f80;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_ROUND_NEAREST: u32 = 0x0000;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_ROUND_DOWN: u32 = 0x2000;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_ROUND_UP: u32 = 0x4000;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_ROUND_TOWARD_ZERO: u32 = 0x6000;
/// See [`_MM_GET_ROUNDING_MODE`](fn._MM_GET_ROUNDING_MODE.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_ROUND_MASK: u32 = 0x6000;
/// See [`_MM_GET_FLUSH_ZERO_MODE`](fn._MM_GET_FLUSH_ZERO_MODE.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_FLUSH_ZERO_MASK: u32 = 0x8000;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_FLUSH_ZERO_ON: u32 = 0x8000;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_FLUSH_ZERO_OFF: u32 = 0x0000;
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_MM_GET_EXCEPTION_MASK)
#[inline]
#[allow(deprecated)] // Deprecated function implemented on top of deprecated function
#[allow(non_snake_case)]
#[target_feature(enable = "sse")]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[deprecated(
since = "1.75.0",
note = "see `_mm_getcsr` documentation - use inline assembly instead"
)]
pub unsafe fn _MM_GET_EXCEPTION_MASK() -> u32 {
_mm_getcsr() & _MM_MASK_MASK
}
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_MM_GET_EXCEPTION_STATE)
#[inline]
#[allow(deprecated)] // Deprecated function implemented on top of deprecated function
#[allow(non_snake_case)]
#[target_feature(enable = "sse")]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[deprecated(
since = "1.75.0",
note = "see `_mm_getcsr` documentation - use inline assembly instead"
)]
pub unsafe fn _MM_GET_EXCEPTION_STATE() -> u32 {
_mm_getcsr() & _MM_EXCEPT_MASK
}
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_MM_GET_FLUSH_ZERO_MODE)
#[inline]
#[allow(deprecated)] // Deprecated function implemented on top of deprecated function
#[allow(non_snake_case)]
#[target_feature(enable = "sse")]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[deprecated(
since = "1.75.0",
note = "see `_mm_getcsr` documentation - use inline assembly instead"
)]
pub unsafe fn _MM_GET_FLUSH_ZERO_MODE() -> u32 {
_mm_getcsr() & _MM_FLUSH_ZERO_MASK
}
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_MM_GET_ROUNDING_MODE)
#[inline]
#[allow(deprecated)] // Deprecated function implemented on top of deprecated function
#[allow(non_snake_case)]
#[target_feature(enable = "sse")]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[deprecated(
since = "1.75.0",
note = "see `_mm_getcsr` documentation - use inline assembly instead"
)]
pub unsafe fn _MM_GET_ROUNDING_MODE() -> u32 {
_mm_getcsr() & _MM_ROUND_MASK
}
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_MM_SET_EXCEPTION_MASK)
#[inline]
#[allow(deprecated)] // Deprecated function implemented on top of deprecated function
#[allow(non_snake_case)]
#[target_feature(enable = "sse")]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[deprecated(
since = "1.75.0",
note = "see `_mm_setcsr` documentation - use inline assembly instead"
)]
pub unsafe fn _MM_SET_EXCEPTION_MASK(x: u32) {
_mm_setcsr((_mm_getcsr() & !_MM_MASK_MASK) | x)
}
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_MM_SET_EXCEPTION_STATE)
#[inline]
#[allow(deprecated)] // Deprecated function implemented on top of deprecated function
#[allow(non_snake_case)]
#[target_feature(enable = "sse")]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[deprecated(
since = "1.75.0",
note = "see `_mm_setcsr` documentation - use inline assembly instead"
)]
pub unsafe fn _MM_SET_EXCEPTION_STATE(x: u32) {
_mm_setcsr((_mm_getcsr() & !_MM_EXCEPT_MASK) | x)
}
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_MM_SET_FLUSH_ZERO_MODE)
#[inline]
#[allow(deprecated)] // Deprecated function implemented on top of deprecated function
#[allow(non_snake_case)]
#[target_feature(enable = "sse")]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[deprecated(
since = "1.75.0",
note = "see `_mm_setcsr` documentation - use inline assembly instead"
)]
pub unsafe fn _MM_SET_FLUSH_ZERO_MODE(x: u32) {
let val = (_mm_getcsr() & !_MM_FLUSH_ZERO_MASK) | x;
// println!("setting csr={:x}", val);
_mm_setcsr(val)
}
/// See [`_mm_setcsr`](fn._mm_setcsr.html)
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_MM_SET_ROUNDING_MODE)
#[inline]
#[allow(deprecated)] // Deprecated function implemented on top of deprecated function
#[allow(non_snake_case)]
#[target_feature(enable = "sse")]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[deprecated(
since = "1.75.0",
note = "see `_mm_setcsr` documentation - use inline assembly instead"
)]
pub unsafe fn _MM_SET_ROUNDING_MODE(x: u32) {
_mm_setcsr((_mm_getcsr() & !_MM_ROUND_MASK) | x)
}
/// See [`_mm_prefetch`](fn._mm_prefetch.html).
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_HINT_T0: i32 = 3;
/// See [`_mm_prefetch`](fn._mm_prefetch.html).
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_HINT_T1: i32 = 2;
/// See [`_mm_prefetch`](fn._mm_prefetch.html).
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_HINT_T2: i32 = 1;
/// See [`_mm_prefetch`](fn._mm_prefetch.html).
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_HINT_NTA: i32 = 0;
/// See [`_mm_prefetch`](fn._mm_prefetch.html).
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_HINT_ET0: i32 = 7;
/// See [`_mm_prefetch`](fn._mm_prefetch.html).
#[stable(feature = "simd_x86", since = "1.27.0")]
pub const _MM_HINT_ET1: i32 = 6;
/// Fetch the cache line that contains address `p` using the given `STRATEGY`.
///
/// The `STRATEGY` must be one of:
///
/// * [`_MM_HINT_T0`](constant._MM_HINT_T0.html): Fetch into all levels of the
/// cache hierarchy.
///
/// * [`_MM_HINT_T1`](constant._MM_HINT_T1.html): Fetch into L2 and higher.
///
/// * [`_MM_HINT_T2`](constant._MM_HINT_T2.html): Fetch into L3 and higher or
/// an implementation-specific choice (e.g., L2 if there is no L3).
///
/// * [`_MM_HINT_NTA`](constant._MM_HINT_NTA.html): Fetch data using the
/// non-temporal access (NTA) hint. It may be a place closer than main memory
/// but outside of the cache hierarchy. This is used to reduce access latency
/// without polluting the cache.
///
/// * [`_MM_HINT_ET0`](constant._MM_HINT_ET0.html) and
/// [`_MM_HINT_ET1`](constant._MM_HINT_ET1.html) are similar to `_MM_HINT_T0`
/// and `_MM_HINT_T1` but indicate an anticipation to write to the address.
///
/// The actual implementation depends on the particular CPU. This instruction
/// is considered a hint, so the CPU is also free to simply ignore the request.
///
/// The amount of prefetched data depends on the cache line size of the
/// specific CPU, but it will be at least 32 bytes.
///
/// Common caveats:
///
/// * Most modern CPUs already automatically prefetch data based on predicted
/// access patterns.
///
/// * Data is usually not fetched if this would cause a TLB miss or a page
/// fault.
///
/// * Too much prefetching can cause unnecessary cache evictions.
///
/// * Prefetching may also fail if there are not enough memory-subsystem
/// resources (e.g., request buffers).
///
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_prefetch)
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(prefetcht0, STRATEGY = _MM_HINT_T0))]
#[cfg_attr(test, assert_instr(prefetcht1, STRATEGY = _MM_HINT_T1))]
#[cfg_attr(test, assert_instr(prefetcht2, STRATEGY = _MM_HINT_T2))]
#[cfg_attr(test, assert_instr(prefetchnta, STRATEGY = _MM_HINT_NTA))]
#[rustc_legacy_const_generics(1)]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_prefetch<const STRATEGY: i32>(p: *const i8) {
static_assert_uimm_bits!(STRATEGY, 3);
// We use the `llvm.prefetch` intrinsic with `cache type` = 1 (data cache).
// `locality` and `rw` are based on our `STRATEGY`.
prefetch(p, (STRATEGY >> 2) & 1, STRATEGY & 3, 1);
}
/// Returns vector of type __m128 with indeterminate elements.
/// Despite being "undefined", this is some valid value and not equivalent to [`mem::MaybeUninit`].
/// In practice, this is equivalent to [`mem::zeroed`].
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_undefined_ps)
#[inline]
#[target_feature(enable = "sse")]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _mm_undefined_ps() -> __m128 {
_mm_set1_ps(0.0)
}
/// Transpose the 4x4 matrix formed by 4 rows of __m128 in place.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_MM_TRANSPOSE4_PS)
#[inline]
#[allow(non_snake_case)]
#[target_feature(enable = "sse")]
#[stable(feature = "simd_x86", since = "1.27.0")]
pub unsafe fn _MM_TRANSPOSE4_PS(
row0: &mut __m128,
row1: &mut __m128,
row2: &mut __m128,
row3: &mut __m128,
) {
let tmp0 = _mm_unpacklo_ps(*row0, *row1);
let tmp2 = _mm_unpacklo_ps(*row2, *row3);
let tmp1 = _mm_unpackhi_ps(*row0, *row1);
let 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);
}
#[allow(improper_ctypes)]
extern "C" {
#[link_name = "llvm.x86.sse.rcp.ss"]
fn rcpss(a: __m128) -> __m128;
#[link_name = "llvm.x86.sse.rcp.ps"]
fn rcpps(a: __m128) -> __m128;
#[link_name = "llvm.x86.sse.rsqrt.ss"]
fn rsqrtss(a: __m128) -> __m128;
#[link_name = "llvm.x86.sse.rsqrt.ps"]
fn rsqrtps(a: __m128) -> __m128;
#[link_name = "llvm.x86.sse.min.ss"]
fn minss(a: __m128, b: __m128) -> __m128;
#[link_name = "llvm.x86.sse.min.ps"]
fn minps(a: __m128, b: __m128) -> __m128;
#[link_name = "llvm.x86.sse.max.ss"]
fn maxss(a: __m128, b: __m128) -> __m128;
#[link_name = "llvm.x86.sse.max.ps"]
fn maxps(a: __m128, b: __m128) -> __m128;
#[link_name = "llvm.x86.sse.cmp.ps"]
fn cmpps(a: __m128, b: __m128, imm8: i8) -> __m128;
#[link_name = "llvm.x86.sse.comieq.ss"]
fn comieq_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.comilt.ss"]
fn comilt_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.comile.ss"]
fn comile_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.comigt.ss"]
fn comigt_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.comige.ss"]
fn comige_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.comineq.ss"]
fn comineq_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.ucomieq.ss"]
fn ucomieq_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.ucomilt.ss"]
fn ucomilt_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.ucomile.ss"]
fn ucomile_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.ucomigt.ss"]
fn ucomigt_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.ucomige.ss"]
fn ucomige_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.ucomineq.ss"]
fn ucomineq_ss(a: __m128, b: __m128) -> i32;
#[link_name = "llvm.x86.sse.cvtss2si"]
fn cvtss2si(a: __m128) -> i32;
#[link_name = "llvm.x86.sse.cvttss2si"]
fn cvttss2si(a: __m128) -> i32;
#[link_name = "llvm.x86.sse.cvtsi2ss"]
fn cvtsi2ss(a: __m128, b: i32) -> __m128;
#[link_name = "llvm.x86.sse.sfence"]
fn sfence();
#[link_name = "llvm.x86.sse.stmxcsr"]
fn stmxcsr(p: *mut i8);
#[link_name = "llvm.x86.sse.ldmxcsr"]
fn ldmxcsr(p: *const i8);
#[link_name = "llvm.prefetch"]
fn prefetch(p: *const i8, rw: i32, loc: i32, ty: i32);
#[link_name = "llvm.x86.sse.cmp.ss"]
fn cmpss(a: __m128, b: __m128, imm8: i8) -> __m128;
}
/// Stores `a` into the memory at `mem_addr` using a non-temporal memory hint.
///
/// `mem_addr` must be aligned on a 16-byte boundary or a general-protection
/// exception _may_ be generated.
///
/// [Intel's documentation](https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_stream_ps)
///
/// # Safety of non-temporal stores
///
/// After using this intrinsic, but before any other access to the memory that this intrinsic
/// mutates, a call to [`_mm_sfence`] must be performed by the thread that used the intrinsic. In
/// particular, functions that call this intrinsic should generally call `_mm_sfence` before they
/// return.
///
/// See [`_mm_sfence`] for details.
#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(movntps))]
#[stable(feature = "simd_x86", since = "1.27.0")]
#[allow(clippy::cast_ptr_alignment)]
pub unsafe fn _mm_stream_ps(mem_addr: *mut f32, a: __m128) {
crate::arch::asm!(
vps!("movntps", ",{a}"),
p = in(reg) mem_addr,
a = in(xmm_reg) a,
options(nostack, preserves_flags),
);
}
#[cfg(test)]
mod tests {
use crate::{hint::black_box, mem::transmute, ptr};
use std::boxed;
use stdarch_test::simd_test;
use crate::core_arch::{simd::*, x86::*};
const NAN: f32 = f32::NAN;
#[simd_test(enable = "sse")]
unsafe fn test_mm_add_ps() {
let a = _mm_setr_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_add_ps(a, b);
assert_eq_m128(r, _mm_setr_ps(-101.0, 25.0, 0.0, -15.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_add_ss() {
let a = _mm_set_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_set_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_add_ss(a, b);
assert_eq_m128(r, _mm_set_ps(-1.0, 5.0, 0.0, -15.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_sub_ps() {
let a = _mm_setr_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_sub_ps(a, b);
assert_eq_m128(r, _mm_setr_ps(99.0, -15.0, 0.0, -5.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_sub_ss() {
let a = _mm_setr_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_sub_ss(a, b);
assert_eq_m128(r, _mm_setr_ps(99.0, 5.0, 0.0, -10.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_mul_ps() {
let a = _mm_setr_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_mul_ps(a, b);
assert_eq_m128(r, _mm_setr_ps(100.0, 100.0, 0.0, 50.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_mul_ss() {
let a = _mm_setr_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_mul_ss(a, b);
assert_eq_m128(r, _mm_setr_ps(100.0, 5.0, 0.0, -10.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_div_ps() {
let a = _mm_setr_ps(-1.0, 5.0, 2.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.2, -5.0);
let r = _mm_div_ps(a, b);
assert_eq_m128(r, _mm_setr_ps(0.01, 0.25, 10.0, 2.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_div_ss() {
let a = _mm_setr_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_div_ss(a, b);
assert_eq_m128(r, _mm_setr_ps(0.01, 5.0, 0.0, -10.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_sqrt_ss() {
let a = _mm_setr_ps(4.0, 13.0, 16.0, 100.0);
let r = _mm_sqrt_ss(a);
let e = _mm_setr_ps(2.0, 13.0, 16.0, 100.0);
assert_eq_m128(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_sqrt_ps() {
let a = _mm_setr_ps(4.0, 13.0, 16.0, 100.0);
let r = _mm_sqrt_ps(a);
let e = _mm_setr_ps(2.0, 3.6055512, 4.0, 10.0);
assert_eq_m128(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_rcp_ss() {
let a = _mm_setr_ps(4.0, 13.0, 16.0, 100.0);
let r = _mm_rcp_ss(a);
let e = _mm_setr_ps(0.24993896, 13.0, 16.0, 100.0);
let rel_err = 0.00048828125;
assert_approx_eq!(get_m128(r, 0), get_m128(e, 0), 2. * rel_err);
for i in 1..4 {
assert_eq!(get_m128(r, i), get_m128(e, i));
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_rcp_ps() {
let a = _mm_setr_ps(4.0, 13.0, 16.0, 100.0);
let r = _mm_rcp_ps(a);
let e = _mm_setr_ps(0.24993896, 0.0769043, 0.06248474, 0.0099983215);
let rel_err = 0.00048828125;
for i in 0..4 {
assert_approx_eq!(get_m128(r, i), get_m128(e, i), 2. * rel_err);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_rsqrt_ss() {
let a = _mm_setr_ps(4.0, 13.0, 16.0, 100.0);
let r = _mm_rsqrt_ss(a);
let e = _mm_setr_ps(0.49987793, 13.0, 16.0, 100.0);
let rel_err = 0.00048828125;
for i in 0..4 {
assert_approx_eq!(get_m128(r, i), get_m128(e, i), 2. * rel_err);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_rsqrt_ps() {
let a = _mm_setr_ps(4.0, 13.0, 16.0, 100.0);
let r = _mm_rsqrt_ps(a);
let e = _mm_setr_ps(0.49987793, 0.2772827, 0.24993896, 0.099990845);
let rel_err = 0.00048828125;
for i in 0..4 {
assert_approx_eq!(get_m128(r, i), get_m128(e, i), 2. * rel_err);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_min_ss() {
let a = _mm_setr_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_min_ss(a, b);
assert_eq_m128(r, _mm_setr_ps(-100.0, 5.0, 0.0, -10.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_min_ps() {
let a = _mm_setr_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_min_ps(a, b);
assert_eq_m128(r, _mm_setr_ps(-100.0, 5.0, 0.0, -10.0));
// `_mm_min_ps` can **not** be implemented using the `simd_min` rust intrinsic. `simd_min`
// is lowered by the llvm codegen backend to `llvm.minnum.v*` llvm intrinsic. This intrinsic
// doesn't specify how -0.0 is handled. Unfortunately it happens to behave different from
// the `minps` x86 instruction on x86. The `llvm.minnum.v*` llvm intrinsic equals
// `r1` to `a` and `r2` to `b`.
let a = _mm_setr_ps(-0.0, 0.0, 0.0, 0.0);
let b = _mm_setr_ps(0.0, 0.0, 0.0, 0.0);
let r1: [u8; 16] = transmute(_mm_min_ps(a, b));
let r2: [u8; 16] = transmute(_mm_min_ps(b, a));
let a: [u8; 16] = transmute(a);
let b: [u8; 16] = transmute(b);
assert_eq!(r1, b);
assert_eq!(r2, a);
assert_ne!(a, b); // sanity check that -0.0 is actually present
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_max_ss() {
let a = _mm_setr_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_max_ss(a, b);
assert_eq_m128(r, _mm_setr_ps(-1.0, 5.0, 0.0, -10.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_max_ps() {
let a = _mm_setr_ps(-1.0, 5.0, 0.0, -10.0);
let b = _mm_setr_ps(-100.0, 20.0, 0.0, -5.0);
let r = _mm_max_ps(a, b);
assert_eq_m128(r, _mm_setr_ps(-1.0, 20.0, 0.0, -5.0));
// Check SSE-specific semantics for -0.0 handling.
let a = _mm_setr_ps(-0.0, 0.0, 0.0, 0.0);
let b = _mm_setr_ps(0.0, 0.0, 0.0, 0.0);
let r1: [u8; 16] = transmute(_mm_max_ps(a, b));
let r2: [u8; 16] = transmute(_mm_max_ps(b, a));
let a: [u8; 16] = transmute(a);
let b: [u8; 16] = transmute(b);
assert_eq!(r1, b);
assert_eq!(r2, a);
assert_ne!(a, b); // sanity check that -0.0 is actually present
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_and_ps() {
let a = transmute(u32x4::splat(0b0011));
let b = transmute(u32x4::splat(0b0101));
let r = _mm_and_ps(*black_box(&a), *black_box(&b));
let e = transmute(u32x4::splat(0b0001));
assert_eq_m128(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_andnot_ps() {
let a = transmute(u32x4::splat(0b0011));
let b = transmute(u32x4::splat(0b0101));
let r = _mm_andnot_ps(*black_box(&a), *black_box(&b));
let e = transmute(u32x4::splat(0b0100));
assert_eq_m128(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_or_ps() {
let a = transmute(u32x4::splat(0b0011));
let b = transmute(u32x4::splat(0b0101));
let r = _mm_or_ps(*black_box(&a), *black_box(&b));
let e = transmute(u32x4::splat(0b0111));
assert_eq_m128(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_xor_ps() {
let a = transmute(u32x4::splat(0b0011));
let b = transmute(u32x4::splat(0b0101));
let r = _mm_xor_ps(*black_box(&a), *black_box(&b));
let e = transmute(u32x4::splat(0b0110));
assert_eq_m128(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpeq_ss() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(-1.0, 5.0, 6.0, 7.0);
let r: u32x4 = transmute(_mm_cmpeq_ss(a, b));
let e: u32x4 = transmute(_mm_setr_ps(f32::from_bits(0), 2.0, 3.0, 4.0));
assert_eq!(r, e);
let b2 = _mm_setr_ps(1.0, 5.0, 6.0, 7.0);
let r2: u32x4 = transmute(_mm_cmpeq_ss(a, b2));
let e2: u32x4 = transmute(_mm_setr_ps(f32::from_bits(0xffffffff), 2.0, 3.0, 4.0));
assert_eq!(r2, e2);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmplt_ss() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(1.0, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = 0u32; // a.extract(0) < b.extract(0)
let c1 = 0u32; // a.extract(0) < c.extract(0)
let d1 = !0u32; // a.extract(0) < d.extract(0)
let rb: u32x4 = transmute(_mm_cmplt_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmplt_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmplt_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmple_ss() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(1.0, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = 0u32; // a.extract(0) <= b.extract(0)
let c1 = !0u32; // a.extract(0) <= c.extract(0)
let d1 = !0u32; // a.extract(0) <= d.extract(0)
let rb: u32x4 = transmute(_mm_cmple_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmple_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmple_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpgt_ss() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(1.0, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = !0u32; // a.extract(0) > b.extract(0)
let c1 = 0u32; // a.extract(0) > c.extract(0)
let d1 = 0u32; // a.extract(0) > d.extract(0)
let rb: u32x4 = transmute(_mm_cmpgt_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmpgt_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmpgt_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpge_ss() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(1.0, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = !0u32; // a.extract(0) >= b.extract(0)
let c1 = !0u32; // a.extract(0) >= c.extract(0)
let d1 = 0u32; // a.extract(0) >= d.extract(0)
let rb: u32x4 = transmute(_mm_cmpge_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmpge_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmpge_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpneq_ss() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(1.0, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = !0u32; // a.extract(0) != b.extract(0)
let c1 = 0u32; // a.extract(0) != c.extract(0)
let d1 = !0u32; // a.extract(0) != d.extract(0)
let rb: u32x4 = transmute(_mm_cmpneq_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmpneq_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmpneq_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpnlt_ss() {
// TODO: this test is exactly the same as for `_mm_cmpge_ss`, but there
// must be a difference. It may have to do with behavior in the
// presence of NaNs (signaling or quiet). If so, we should add tests
// for those.
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(1.0, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = !0u32; // a.extract(0) >= b.extract(0)
let c1 = !0u32; // a.extract(0) >= c.extract(0)
let d1 = 0u32; // a.extract(0) >= d.extract(0)
let rb: u32x4 = transmute(_mm_cmpnlt_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmpnlt_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmpnlt_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpnle_ss() {
// TODO: this test is exactly the same as for `_mm_cmpgt_ss`, but there
// must be a difference. It may have to do with behavior in the
// presence
// of NaNs (signaling or quiet). If so, we should add tests for those.
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(1.0, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = !0u32; // a.extract(0) > b.extract(0)
let c1 = 0u32; // a.extract(0) > c.extract(0)
let d1 = 0u32; // a.extract(0) > d.extract(0)
let rb: u32x4 = transmute(_mm_cmpnle_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmpnle_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmpnle_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpngt_ss() {
// TODO: this test is exactly the same as for `_mm_cmple_ss`, but there
// must be a difference. It may have to do with behavior in the
// presence of NaNs (signaling or quiet). If so, we should add tests
// for those.
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(1.0, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = 0u32; // a.extract(0) <= b.extract(0)
let c1 = !0u32; // a.extract(0) <= c.extract(0)
let d1 = !0u32; // a.extract(0) <= d.extract(0)
let rb: u32x4 = transmute(_mm_cmpngt_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmpngt_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmpngt_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpnge_ss() {
// TODO: this test is exactly the same as for `_mm_cmplt_ss`, but there
// must be a difference. It may have to do with behavior in the
// presence of NaNs (signaling or quiet). If so, we should add tests
// for those.
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(1.0, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = 0u32; // a.extract(0) < b.extract(0)
let c1 = 0u32; // a.extract(0) < c.extract(0)
let d1 = !0u32; // a.extract(0) < d.extract(0)
let rb: u32x4 = transmute(_mm_cmpnge_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmpnge_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmpnge_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpord_ss() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(NAN, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = !0u32; // a.extract(0) ord b.extract(0)
let c1 = 0u32; // a.extract(0) ord c.extract(0)
let d1 = !0u32; // a.extract(0) ord d.extract(0)
let rb: u32x4 = transmute(_mm_cmpord_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmpord_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmpord_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpunord_ss() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(0.0, 5.0, 6.0, 7.0);
let c = _mm_setr_ps(NAN, 5.0, 6.0, 7.0);
let d = _mm_setr_ps(2.0, 5.0, 6.0, 7.0);
let b1 = 0u32; // a.extract(0) unord b.extract(0)
let c1 = !0u32; // a.extract(0) unord c.extract(0)
let d1 = 0u32; // a.extract(0) unord d.extract(0)
let rb: u32x4 = transmute(_mm_cmpunord_ss(a, b));
let eb: u32x4 = transmute(_mm_setr_ps(f32::from_bits(b1), 2.0, 3.0, 4.0));
assert_eq!(rb, eb);
let rc: u32x4 = transmute(_mm_cmpunord_ss(a, c));
let ec: u32x4 = transmute(_mm_setr_ps(f32::from_bits(c1), 2.0, 3.0, 4.0));
assert_eq!(rc, ec);
let rd: u32x4 = transmute(_mm_cmpunord_ss(a, d));
let ed: u32x4 = transmute(_mm_setr_ps(f32::from_bits(d1), 2.0, 3.0, 4.0));
assert_eq!(rd, ed);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpeq_ps() {
let a = _mm_setr_ps(10.0, 50.0, 1.0, NAN);
let b = _mm_setr_ps(15.0, 20.0, 1.0, NAN);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(fls, fls, tru, fls);
let r: u32x4 = transmute(_mm_cmpeq_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmplt_ps() {
let a = _mm_setr_ps(10.0, 50.0, 1.0, NAN);
let b = _mm_setr_ps(15.0, 20.0, 1.0, NAN);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(tru, fls, fls, fls);
let r: u32x4 = transmute(_mm_cmplt_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmple_ps() {
let a = _mm_setr_ps(10.0, 50.0, 1.0, 4.0);
let b = _mm_setr_ps(15.0, 20.0, 1.0, NAN);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(tru, fls, tru, fls);
let r: u32x4 = transmute(_mm_cmple_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpgt_ps() {
let a = _mm_setr_ps(10.0, 50.0, 1.0, NAN);
let b = _mm_setr_ps(15.0, 20.0, 1.0, 42.0);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(fls, tru, fls, fls);
let r: u32x4 = transmute(_mm_cmpgt_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpge_ps() {
let a = _mm_setr_ps(10.0, 50.0, 1.0, NAN);
let b = _mm_setr_ps(15.0, 20.0, 1.0, 42.0);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(fls, tru, tru, fls);
let r: u32x4 = transmute(_mm_cmpge_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpneq_ps() {
let a = _mm_setr_ps(10.0, 50.0, 1.0, NAN);
let b = _mm_setr_ps(15.0, 20.0, 1.0, NAN);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(tru, tru, fls, tru);
let r: u32x4 = transmute(_mm_cmpneq_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpnlt_ps() {
let a = _mm_setr_ps(10.0, 50.0, 1.0, NAN);
let b = _mm_setr_ps(15.0, 20.0, 1.0, 5.0);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(fls, tru, tru, tru);
let r: u32x4 = transmute(_mm_cmpnlt_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpnle_ps() {
let a = _mm_setr_ps(10.0, 50.0, 1.0, NAN);
let b = _mm_setr_ps(15.0, 20.0, 1.0, 5.0);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(fls, tru, fls, tru);
let r: u32x4 = transmute(_mm_cmpnle_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpngt_ps() {
let a = _mm_setr_ps(10.0, 50.0, 1.0, NAN);
let b = _mm_setr_ps(15.0, 20.0, 1.0, 5.0);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(tru, fls, tru, tru);
let r: u32x4 = transmute(_mm_cmpngt_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpnge_ps() {
let a = _mm_setr_ps(10.0, 50.0, 1.0, NAN);
let b = _mm_setr_ps(15.0, 20.0, 1.0, 5.0);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(tru, fls, fls, tru);
let r: u32x4 = transmute(_mm_cmpnge_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpord_ps() {
let a = _mm_setr_ps(10.0, 50.0, NAN, NAN);
let b = _mm_setr_ps(15.0, NAN, 1.0, NAN);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(tru, fls, fls, fls);
let r: u32x4 = transmute(_mm_cmpord_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cmpunord_ps() {
let a = _mm_setr_ps(10.0, 50.0, NAN, NAN);
let b = _mm_setr_ps(15.0, NAN, 1.0, NAN);
let tru = !0u32;
let fls = 0u32;
let e = u32x4::new(fls, tru, tru, tru);
let r: u32x4 = transmute(_mm_cmpunord_ps(a, b));
assert_eq!(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_comieq_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[1i32, 0, 0, 0];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_comieq_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_comieq_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_comilt_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[0i32, 1, 0, 0];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_comilt_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_comilt_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_comile_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[1i32, 1, 0, 0];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_comile_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_comile_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_comigt_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[1i32, 0, 1, 0];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_comige_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_comige_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_comineq_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[0i32, 1, 1, 1];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_comineq_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_comineq_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_ucomieq_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[1i32, 0, 0, 0];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_ucomieq_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_ucomieq_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_ucomilt_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[0i32, 1, 0, 0];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_ucomilt_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_ucomilt_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_ucomile_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[1i32, 1, 0, 0];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_ucomile_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_ucomile_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_ucomigt_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[0i32, 0, 1, 0];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_ucomigt_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_ucomigt_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_ucomige_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[1i32, 0, 1, 0];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_ucomige_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_ucomige_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_ucomineq_ss() {
let aa = &[3.0f32, 12.0, 23.0, NAN];
let bb = &[3.0f32, 47.5, 1.5, NAN];
let ee = &[0i32, 1, 1, 1];
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
let r = _mm_ucomineq_ss(a, b);
assert_eq!(
ee[i], r,
"_mm_ucomineq_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r, ee[i], i
);
}
}
#[allow(deprecated)] // FIXME: This test uses deprecated CSR access functions
#[simd_test(enable = "sse")]
#[cfg_attr(miri, ignore)] // Uses _mm_setcsr, which is not supported by Miri
unsafe fn test_mm_comieq_ss_vs_ucomieq_ss() {
// If one of the arguments is a quiet NaN `comieq_ss` should signal an
// Invalid Operation Exception while `ucomieq_ss` should not.
let aa = &[3.0f32, NAN, 23.0, NAN];
let bb = &[3.0f32, 47.5, NAN, NAN];
let ee = &[1i32, 0, 0, 0];
let exc = &[0u32, 1, 1, 1]; // Should comieq_ss signal an exception?
for i in 0..4 {
let a = _mm_setr_ps(aa[i], 1.0, 2.0, 3.0);
let b = _mm_setr_ps(bb[i], 0.0, 2.0, 4.0);
_MM_SET_EXCEPTION_STATE(0);
let r1 = _mm_comieq_ss(*black_box(&a), b);
let s1 = _MM_GET_EXCEPTION_STATE();
_MM_SET_EXCEPTION_STATE(0);
let r2 = _mm_ucomieq_ss(*black_box(&a), b);
let s2 = _MM_GET_EXCEPTION_STATE();
assert_eq!(
ee[i], r1,
"_mm_comeq_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r1, ee[i], i
);
assert_eq!(
ee[i], r2,
"_mm_ucomeq_ss({:?}, {:?}) = {}, expected: {} (i={})",
a, b, r2, ee[i], i
);
assert_eq!(
s1,
exc[i] * _MM_EXCEPT_INVALID,
"_mm_comieq_ss() set exception flags: {} (i={})",
s1,
i
);
assert_eq!(
s2,
0, // ucomieq_ss should not signal an exception
"_mm_ucomieq_ss() set exception flags: {} (i={})",
s2,
i
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cvtss_si32() {
let inputs = &[42.0f32, -3.1, 4.0e10, 4.0e-20, NAN, 2147483500.1];
let result = &[42i32, -3, i32::MIN, 0, i32::MIN, 2147483520];
for i in 0..inputs.len() {
let x = _mm_setr_ps(inputs[i], 1.0, 3.0, 4.0);
let e = result[i];
let r = _mm_cvtss_si32(x);
assert_eq!(
e, r,
"TestCase #{} _mm_cvtss_si32({:?}) = {}, expected: {}",
i, x, r, e
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cvttss_si32() {
let inputs = &[
(42.0f32, 42i32),
(-31.4, -31),
(-33.5, -33),
(-34.5, -34),
(10.999, 10),
(-5.99, -5),
(4.0e10, i32::MIN),
(4.0e-10, 0),
(NAN, i32::MIN),
(2147483500.1, 2147483520),
];
for (i, &(xi, e)) in inputs.iter().enumerate() {
let x = _mm_setr_ps(xi, 1.0, 3.0, 4.0);
let r = _mm_cvttss_si32(x);
assert_eq!(
e, r,
"TestCase #{} _mm_cvttss_si32({:?}) = {}, expected: {}",
i, x, r, e
);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cvtsi32_ss() {
let inputs = &[
(4555i32, 4555.0f32),
(322223333, 322223330.0),
(-432, -432.0),
(-322223333, -322223330.0),
];
for &(x, f) in inputs.iter() {
let a = _mm_setr_ps(5.0, 6.0, 7.0, 8.0);
let r = _mm_cvtsi32_ss(a, x);
let e = _mm_setr_ps(f, 6.0, 7.0, 8.0);
assert_eq_m128(e, r);
}
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_cvtss_f32() {
let a = _mm_setr_ps(312.0134, 5.0, 6.0, 7.0);
assert_eq!(_mm_cvtss_f32(a), 312.0134);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_set_ss() {
let r = _mm_set_ss(black_box(4.25));
assert_eq_m128(r, _mm_setr_ps(4.25, 0.0, 0.0, 0.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_set1_ps() {
let r1 = _mm_set1_ps(black_box(4.25));
let r2 = _mm_set_ps1(black_box(4.25));
assert_eq!(get_m128(r1, 0), 4.25);
assert_eq!(get_m128(r1, 1), 4.25);
assert_eq!(get_m128(r1, 2), 4.25);
assert_eq!(get_m128(r1, 3), 4.25);
assert_eq!(get_m128(r2, 0), 4.25);
assert_eq!(get_m128(r2, 1), 4.25);
assert_eq!(get_m128(r2, 2), 4.25);
assert_eq!(get_m128(r2, 3), 4.25);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_set_ps() {
let r = _mm_set_ps(
black_box(1.0),
black_box(2.0),
black_box(3.0),
black_box(4.0),
);
assert_eq!(get_m128(r, 0), 4.0);
assert_eq!(get_m128(r, 1), 3.0);
assert_eq!(get_m128(r, 2), 2.0);
assert_eq!(get_m128(r, 3), 1.0);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_setr_ps() {
let r = _mm_setr_ps(
black_box(1.0),
black_box(2.0),
black_box(3.0),
black_box(4.0),
);
assert_eq_m128(r, _mm_setr_ps(1.0, 2.0, 3.0, 4.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_setzero_ps() {
let r = *black_box(&_mm_setzero_ps());
assert_eq_m128(r, _mm_set1_ps(0.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_shuffle() {
assert_eq!(_MM_SHUFFLE(0, 1, 1, 3), 0b00_01_01_11);
assert_eq!(_MM_SHUFFLE(3, 1, 1, 0), 0b11_01_01_00);
assert_eq!(_MM_SHUFFLE(1, 2, 2, 1), 0b01_10_10_01);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_shuffle_ps() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(5.0, 6.0, 7.0, 8.0);
let r = _mm_shuffle_ps::<0b00_01_01_11>(a, b);
assert_eq_m128(r, _mm_setr_ps(4.0, 2.0, 6.0, 5.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_unpackhi_ps() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(5.0, 6.0, 7.0, 8.0);
let r = _mm_unpackhi_ps(a, b);
assert_eq_m128(r, _mm_setr_ps(3.0, 7.0, 4.0, 8.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_unpacklo_ps() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(5.0, 6.0, 7.0, 8.0);
let r = _mm_unpacklo_ps(a, b);
assert_eq_m128(r, _mm_setr_ps(1.0, 5.0, 2.0, 6.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_movehl_ps() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(5.0, 6.0, 7.0, 8.0);
let r = _mm_movehl_ps(a, b);
assert_eq_m128(r, _mm_setr_ps(7.0, 8.0, 3.0, 4.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_movelh_ps() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(5.0, 6.0, 7.0, 8.0);
let r = _mm_movelh_ps(a, b);
assert_eq_m128(r, _mm_setr_ps(1.0, 2.0, 5.0, 6.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_load_ss() {
let a = 42.0f32;
let r = _mm_load_ss(ptr::addr_of!(a));
assert_eq_m128(r, _mm_setr_ps(42.0, 0.0, 0.0, 0.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_load1_ps() {
let a = 42.0f32;
let r = _mm_load1_ps(ptr::addr_of!(a));
assert_eq_m128(r, _mm_setr_ps(42.0, 42.0, 42.0, 42.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_load_ps() {
let vals = &[1.0f32, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0];
let mut p = vals.as_ptr();
let mut fixup = 0.0f32;
// Make sure p is aligned, otherwise we might get a
// (signal: 11, SIGSEGV: invalid memory reference)
let unalignment = (p as usize) & 0xf;
if unalignment != 0 {
let delta = (16 - unalignment) >> 2;
fixup = delta as f32;
p = p.add(delta);
}
let r = _mm_load_ps(p);
let e = _mm_add_ps(_mm_setr_ps(1.0, 2.0, 3.0, 4.0), _mm_set1_ps(fixup));
assert_eq_m128(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_loadu_ps() {
let vals = &[1.0f32, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0];
let p = vals.as_ptr().add(3);
let r = _mm_loadu_ps(black_box(p));
assert_eq_m128(r, _mm_setr_ps(4.0, 5.0, 6.0, 7.0));
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_loadr_ps() {
let vals = &[1.0f32, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0];
let mut p = vals.as_ptr();
let mut fixup = 0.0f32;
// Make sure p is aligned, otherwise we might get a
// (signal: 11, SIGSEGV: invalid memory reference)
let unalignment = (p as usize) & 0xf;
if unalignment != 0 {
let delta = (16 - unalignment) >> 2;
fixup = delta as f32;
p = p.add(delta);
}
let r = _mm_loadr_ps(p);
let e = _mm_add_ps(_mm_setr_ps(4.0, 3.0, 2.0, 1.0), _mm_set1_ps(fixup));
assert_eq_m128(r, e);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_store_ss() {
let mut vals = [0.0f32; 8];
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
_mm_store_ss(vals.as_mut_ptr().add(1), a);
assert_eq!(vals[0], 0.0);
assert_eq!(vals[1], 1.0);
assert_eq!(vals[2], 0.0);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_store1_ps() {
let mut vals = [0.0f32; 8];
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let mut ofs = 0;
let mut p = vals.as_mut_ptr();
if (p as usize) & 0xf != 0 {
ofs = (16 - ((p as usize) & 0xf)) >> 2;
p = p.add(ofs);
}
_mm_store1_ps(p, *black_box(&a));
if ofs > 0 {
assert_eq!(vals[ofs - 1], 0.0);
}
assert_eq!(vals[ofs + 0], 1.0);
assert_eq!(vals[ofs + 1], 1.0);
assert_eq!(vals[ofs + 2], 1.0);
assert_eq!(vals[ofs + 3], 1.0);
assert_eq!(vals[ofs + 4], 0.0);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_store_ps() {
let mut vals = [0.0f32; 8];
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let mut ofs = 0;
let mut p = vals.as_mut_ptr();
// Align p to 16-byte boundary
if (p as usize) & 0xf != 0 {
ofs = (16 - ((p as usize) & 0xf)) >> 2;
p = p.add(ofs);
}
_mm_store_ps(p, *black_box(&a));
if ofs > 0 {
assert_eq!(vals[ofs - 1], 0.0);
}
assert_eq!(vals[ofs + 0], 1.0);
assert_eq!(vals[ofs + 1], 2.0);
assert_eq!(vals[ofs + 2], 3.0);
assert_eq!(vals[ofs + 3], 4.0);
assert_eq!(vals[ofs + 4], 0.0);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_storer_ps() {
let mut vals = [0.0f32; 8];
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let mut ofs = 0;
let mut p = vals.as_mut_ptr();
// Align p to 16-byte boundary
if (p as usize) & 0xf != 0 {
ofs = (16 - ((p as usize) & 0xf)) >> 2;
p = p.add(ofs);
}
_mm_storer_ps(p, *black_box(&a));
if ofs > 0 {
assert_eq!(vals[ofs - 1], 0.0);
}
assert_eq!(vals[ofs + 0], 4.0);
assert_eq!(vals[ofs + 1], 3.0);
assert_eq!(vals[ofs + 2], 2.0);
assert_eq!(vals[ofs + 3], 1.0);
assert_eq!(vals[ofs + 4], 0.0);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_storeu_ps() {
let mut vals = [0.0f32; 8];
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let mut ofs = 0;
let mut p = vals.as_mut_ptr();
// Make sure p is **not** aligned to 16-byte boundary
if (p as usize) & 0xf == 0 {
ofs = 1;
p = p.add(1);
}
_mm_storeu_ps(p, *black_box(&a));
if ofs > 0 {
assert_eq!(vals[ofs - 1], 0.0);
}
assert_eq!(vals[ofs + 0], 1.0);
assert_eq!(vals[ofs + 1], 2.0);
assert_eq!(vals[ofs + 2], 3.0);
assert_eq!(vals[ofs + 3], 4.0);
assert_eq!(vals[ofs + 4], 0.0);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_move_ss() {
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let b = _mm_setr_ps(5.0, 6.0, 7.0, 8.0);
let r = _mm_move_ss(a, b);
let e = _mm_setr_ps(5.0, 2.0, 3.0, 4.0);
assert_eq_m128(e, r);
}
#[simd_test(enable = "sse")]
unsafe fn test_mm_movemask_ps() {
let r = _mm_movemask_ps(_mm_setr_ps(-1.0, 5.0, -5.0, 0.0));
assert_eq!(r, 0b0101);
let r = _mm_movemask_ps(_mm_setr_ps(-1.0, -5.0, -5.0, 0.0));
assert_eq!(r, 0b0111);
}
#[simd_test(enable = "sse")]
// Miri cannot support this until it is clear how it fits in the Rust memory model
#[cfg_attr(miri, ignore)]
unsafe fn test_mm_sfence() {
_mm_sfence();
}
#[allow(deprecated)] // FIXME: This tests functions that are immediate UB
#[simd_test(enable = "sse")]
#[cfg_attr(miri, ignore)] // Miri does not support accesing the CSR
unsafe fn test_mm_getcsr_setcsr_1() {
let saved_csr = _mm_getcsr();
let a = _mm_setr_ps(1.1e-36, 0.0, 0.0, 1.0);
let b = _mm_setr_ps(0.001, 0.0, 0.0, 1.0);
_MM_SET_FLUSH_ZERO_MODE(_MM_FLUSH_ZERO_ON);
let r = _mm_mul_ps(*black_box(&a), *black_box(&b));
_mm_setcsr(saved_csr);
let exp = _mm_setr_ps(0.0, 0.0, 0.0, 1.0);
assert_eq_m128(r, exp); // first component is a denormalized f32
}
#[allow(deprecated)] // FIXME: This tests functions that are immediate UB
#[simd_test(enable = "sse")]
#[cfg_attr(miri, ignore)] // Miri does not support accesing the CSR
unsafe fn test_mm_getcsr_setcsr_2() {
// Same as _mm_setcsr_1 test, but with opposite flag value.
let saved_csr = _mm_getcsr();
let a = _mm_setr_ps(1.1e-36, 0.0, 0.0, 1.0);
let b = _mm_setr_ps(0.001, 0.0, 0.0, 1.0);
_MM_SET_FLUSH_ZERO_MODE(_MM_FLUSH_ZERO_OFF);
let r = _mm_mul_ps(*black_box(&a), *black_box(&b));
_mm_setcsr(saved_csr);
let exp = _mm_setr_ps(1.1e-39, 0.0, 0.0, 1.0);
assert_eq_m128(r, exp); // first component is a denormalized f32
}
#[allow(deprecated)] // FIXME: This tests functions that are immediate UB
#[simd_test(enable = "sse")]
#[cfg_attr(miri, ignore)] // Miri does not support accesing the CSR
unsafe fn test_mm_getcsr_setcsr_underflow() {
_MM_SET_EXCEPTION_STATE(0);
let a = _mm_setr_ps(1.1e-36, 0.0, 0.0, 1.0);
let b = _mm_setr_ps(1e-5, 0.0, 0.0, 1.0);
assert_eq!(_MM_GET_EXCEPTION_STATE(), 0); // just to be sure
let r = _mm_mul_ps(*black_box(&a), *black_box(&b));
let exp = _mm_setr_ps(1.1e-41, 0.0, 0.0, 1.0);
assert_eq_m128(r, exp);
let underflow = _MM_GET_EXCEPTION_STATE() & _MM_EXCEPT_UNDERFLOW != 0;
assert!(underflow);
}
#[simd_test(enable = "sse")]
unsafe fn test_MM_TRANSPOSE4_PS() {
let mut a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);
let mut b = _mm_setr_ps(5.0, 6.0, 7.0, 8.0);
let mut c = _mm_setr_ps(9.0, 10.0, 11.0, 12.0);
let mut d = _mm_setr_ps(13.0, 14.0, 15.0, 16.0);
_MM_TRANSPOSE4_PS(&mut a, &mut b, &mut c, &mut d);
assert_eq_m128(a, _mm_setr_ps(1.0, 5.0, 9.0, 13.0));
assert_eq_m128(b, _mm_setr_ps(2.0, 6.0, 10.0, 14.0));
assert_eq_m128(c, _mm_setr_ps(3.0, 7.0, 11.0, 15.0));
assert_eq_m128(d, _mm_setr_ps(4.0, 8.0, 12.0, 16.0));
}
#[repr(align(16))]
struct Memory {
pub data: [f32; 4],
}
#[simd_test(enable = "sse")]
// Miri cannot support this until it is clear how it fits in the Rust memory model
// (non-temporal store)
#[cfg_attr(miri, ignore)]
unsafe fn test_mm_stream_ps() {
let a = _mm_set1_ps(7.0);
let mut mem = Memory { data: [-1.0; 4] };
_mm_stream_ps(ptr::addr_of_mut!(mem.data[0]), a);
for i in 0..4 {
assert_eq!(mem.data[i], get_m128(a, i));
}
}
}