aboutsummaryrefslogtreecommitdiff
path: root/sysdeps/ia64/fpu/s_cos.S
blob: cd715b4d22c839f9c5f210a78949e23fae12b376 (plain)
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
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
.file "sincos.s"

// Copyright (c) 2000, 2001, Intel Corporation
// All rights reserved.
//
// Contributed 2/2/2000 by John Harrison, Ted Kubaska, Bob Norin, Shane Story,
// and Ping Tak Peter Tang of the Computational Software Lab, Intel Corporation.
//
// WARRANTY DISCLAIMER
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL INTEL OR ITS
// CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
// EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
// PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
// OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY OR TORT (INCLUDING
// NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
// SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
//
// Intel Corporation is the author of this code, and requests that all
// problem reports or change requests be submitted to it directly at
// http://developer.intel.com/opensource.
//
// History
//==============================================================
// 2/02/00  Initial revision
// 4/02/00  Unwind support added.
// 6/16/00  Updated tables to enforce symmetry
// 8/31/00  Saved 2 cycles in main path, and 9 in other paths.
// 9/20/00  The updated tables regressed to an old version, so reinstated them
// 10/18/00 Changed one table entry to ensure symmetry
// 1/03/01  Improved speed, fixed flag settings for small arguments.

// API
//==============================================================
// double sin( double x);
// double cos( double x);
//
// Overview of operation
//==============================================================
//
// Step 1
// ======
// Reduce x to region -1/2*pi/2^k ===== 0 ===== +1/2*pi/2^k  where k=4
//    divide x by pi/2^k.
//    Multiply by 2^k/pi.
//    nfloat = Round result to integer (round-to-nearest)
//
// r = x -  nfloat * pi/2^k
//    Do this as (x -  nfloat * HIGH(pi/2^k)) - nfloat * LOW(pi/2^k) for increased accuracy.
//    pi/2^k is stored as two numbers that when added make pi/2^k.
//       pi/2^k = HIGH(pi/2^k) + LOW(pi/2^k)
//
// x = (nfloat * pi/2^k) + r
//    r is small enough that we can use a polynomial approximation
//    and is referred to as the reduced argument.
//
// Step 3
// ======
// Take the unreduced part and remove the multiples of 2pi.
// So nfloat = nfloat (with lower k+1 bits cleared) + lower k+1 bits
//
//    nfloat (with lower k+1 bits cleared) is a multiple of 2^(k+1)
//    N * 2^(k+1)
//    nfloat * pi/2^k = N * 2^(k+1) * pi/2^k + (lower k+1 bits) * pi/2^k
//    nfloat * pi/2^k = N * 2 * pi + (lower k+1 bits) * pi/2^k
//    nfloat * pi/2^k = N2pi + M * pi/2^k
//
//
// Sin(x) = Sin((nfloat * pi/2^k) + r)
//        = Sin(nfloat * pi/2^k) * Cos(r) + Cos(nfloat * pi/2^k) * Sin(r)
//
//          Sin(nfloat * pi/2^k) = Sin(N2pi + Mpi/2^k)
//                               = Sin(N2pi)Cos(Mpi/2^k) + Cos(N2pi)Sin(Mpi/2^k)
//                               = Sin(Mpi/2^k)
//
//          Cos(nfloat * pi/2^k) = Cos(N2pi + Mpi/2^k)
//                               = Cos(N2pi)Cos(Mpi/2^k) + Sin(N2pi)Sin(Mpi/2^k)
//                               = Cos(Mpi/2^k)
//
// Sin(x) = Sin(Mpi/2^k) Cos(r) + Cos(Mpi/2^k) Sin(r)
//
//
// Step 4
// ======
// 0 <= M < 2^(k+1)
// There are 2^(k+1) Sin entries in a table.
// There are 2^(k+1) Cos entries in a table.
//
// Get Sin(Mpi/2^k) and Cos(Mpi/2^k) by table lookup.
//
//
// Step 5
// ======
// Calculate Cos(r) and Sin(r) by polynomial approximation.
//
// Cos(r) = 1 + r^2 q1  + r^4 q2 + r^6 q3 + ... = Series for Cos
// Sin(r) = r + r^3 p1  + r^5 p2 + r^7 p3 + ... = Series for Sin
//
// and the coefficients q1, q2, ... and p1, p2, ... are stored in a table
//
//
// Calculate
// Sin(x) = Sin(Mpi/2^k) Cos(r) + Cos(Mpi/2^k) Sin(r)
//
// as follows
//
//    Sm = Sin(Mpi/2^k) and Cm = Cos(Mpi/2^k)
//    rsq = r*r
//
//
//    P = p1 + r^2p2 + r^4p3 + r^6p4
//    Q = q1 + r^2q2 + r^4q3 + r^6q4
//
//       rcub = r * rsq
//       Sin(r) = r + rcub * P
//              = r + r^3p1  + r^5p2 + r^7p3 + r^9p4 + ... = Sin(r)
//
//            The coefficients are not exactly these values, but almost.
//
//            p1 = -1/6  = -1/3!
//            p2 = 1/120 =  1/5!
//            p3 = -1/5040 = -1/7!
//            p4 = 1/362889 = 1/9!
//
//       P =  r + rcub * P
//
//    Answer = Sm Cos(r) + Cm P
//
//       Cos(r) = 1 + rsq Q
//       Cos(r) = 1 + r^2 Q
//       Cos(r) = 1 + r^2 (q1 + r^2q2 + r^4q3 + r^6q4)
//       Cos(r) = 1 + r^2q1 + r^4q2 + r^6q3 + r^8q4 + ...
//
//       Sm Cos(r) = Sm(1 + rsq Q)
//       Sm Cos(r) = Sm + Sm rsq Q
//       Sm Cos(r) = Sm + s_rsq Q
//       Q         = Sm + s_rsq Q
//
// Then,
//
//    Answer = Q + Cm P

#include "libm_support.h"

// Registers used
//==============================================================
// general input registers:
// r14 -> r19
// r32 -> r45

// predicate registers used:
// p6 -> p14

// floating-point registers used
// f9 -> f15
// f32 -> f61

// Assembly macros
//==============================================================
sind_NORM_f8                 = f9
sind_W                       = f10
sind_int_Nfloat              = f11
sind_Nfloat                  = f12

sind_r                       = f13
sind_rsq                     = f14
sind_rcub                    = f15

sind_Inv_Pi_by_16            = f32
sind_Pi_by_16_hi             = f33
sind_Pi_by_16_lo             = f34

sind_Inv_Pi_by_64            = f35
sind_Pi_by_64_hi             = f36
sind_Pi_by_64_lo             = f37

sind_Sm                      = f38
sind_Cm                      = f39

sind_P1                      = f40
sind_Q1                      = f41
sind_P2                      = f42
sind_Q2                      = f43
sind_P3                      = f44
sind_Q3                      = f45
sind_P4                      = f46
sind_Q4                      = f47

sind_P_temp1                 = f48
sind_P_temp2                 = f49

sind_Q_temp1                 = f50
sind_Q_temp2                 = f51

sind_P                       = f52
sind_Q                       = f53

sind_srsq                    = f54

sind_SIG_INV_PI_BY_16_2TO61  = f55
sind_RSHF_2TO61              = f56
sind_RSHF                    = f57
sind_2TOM61                  = f58
sind_NFLOAT                  = f59
sind_W_2TO61_RSH             = f60

fp_tmp                       = f61

/////////////////////////////////////////////////////////////

sind_AD_1                    = r33
sind_AD_2                    = r34
sind_exp_limit               = r35
sind_r_signexp               = r36
sind_AD_beta_table           = r37
sind_r_sincos                = r38

sind_r_exp                   = r39
sind_r_17_ones               = r40

sind_GR_sig_inv_pi_by_16     = r14
sind_GR_rshf_2to61           = r15
sind_GR_rshf                 = r16
sind_GR_exp_2tom61           = r17
sind_GR_n                    = r18
sind_GR_m                    = r19
sind_GR_32m                  = r19

gr_tmp                       = r41
GR_SAVE_PFS                  = r41
GR_SAVE_B0                   = r42
GR_SAVE_GP                   = r43


#ifdef _LIBC
.rodata
#else
.data
#endif

.align 16
double_sind_pi:
ASM_TYPE_DIRECTIVE(double_sind_pi,@object)
//   data8 0xA2F9836E4E44152A, 0x00004001 // 16/pi (significand loaded w/ setf)
//         c90fdaa22168c234
   data8 0xC90FDAA22168C234, 0x00003FFC // pi/16 hi
//         c4c6628b80dc1cd1  29024e088a
   data8 0xC4C6628B80DC1CD1, 0x00003FBC // pi/16 lo
ASM_SIZE_DIRECTIVE(double_sind_pi)

double_sind_pq_k4:
ASM_TYPE_DIRECTIVE(double_sind_pq_k4,@object)
   data8 0x3EC71C963717C63A // P4
   data8 0x3EF9FFBA8F191AE6 // Q4
   data8 0xBF2A01A00F4E11A8 // P3
   data8 0xBF56C16C05AC77BF // Q3
   data8 0x3F8111111110F167 // P2
   data8 0x3FA555555554DD45 // Q2
   data8 0xBFC5555555555555 // P1
   data8 0xBFDFFFFFFFFFFFFC // Q1
ASM_SIZE_DIRECTIVE(double_sind_pq_k4)


double_sin_cos_beta_k4:
ASM_TYPE_DIRECTIVE(double_sin_cos_beta_k4,@object)
data8 0x0000000000000000 , 0x00000000 // sin( 0 pi/16)  S0
data8 0x8000000000000000 , 0x00003fff // cos( 0 pi/16)  C0

data8 0xc7c5c1e34d3055b3 , 0x00003ffc // sin( 1 pi/16)  S1
data8 0xfb14be7fbae58157 , 0x00003ffe // cos( 1 pi/16)  C1

data8 0xc3ef1535754b168e , 0x00003ffd // sin( 2 pi/16)  S2
data8 0xec835e79946a3146 , 0x00003ffe // cos( 2 pi/16)  C2

data8 0x8e39d9cd73464364 , 0x00003ffe // sin( 3 pi/16)  S3
data8 0xd4db3148750d181a , 0x00003ffe // cos( 3 pi/16)  C3

data8 0xb504f333f9de6484 , 0x00003ffe // sin( 4 pi/16)  S4
data8 0xb504f333f9de6484 , 0x00003ffe // cos( 4 pi/16)  C4


data8 0xd4db3148750d181a , 0x00003ffe // sin( 5 pi/16)  C3
data8 0x8e39d9cd73464364 , 0x00003ffe // cos( 5 pi/16)  S3

data8 0xec835e79946a3146 , 0x00003ffe // sin( 6 pi/16)  C2
data8 0xc3ef1535754b168e , 0x00003ffd // cos( 6 pi/16)  S2

data8 0xfb14be7fbae58157 , 0x00003ffe // sin( 7 pi/16)  C1
data8 0xc7c5c1e34d3055b3 , 0x00003ffc // cos( 7 pi/16)  S1

data8 0x8000000000000000 , 0x00003fff // sin( 8 pi/16)  C0
data8 0x0000000000000000 , 0x00000000 // cos( 8 pi/16)  S0


data8 0xfb14be7fbae58157 , 0x00003ffe // sin( 9 pi/16)  C1
data8 0xc7c5c1e34d3055b3 , 0x0000bffc // cos( 9 pi/16)  -S1

data8 0xec835e79946a3146 , 0x00003ffe // sin(10 pi/16)  C2
data8 0xc3ef1535754b168e , 0x0000bffd // cos(10 pi/16)  -S2

data8 0xd4db3148750d181a , 0x00003ffe // sin(11 pi/16)  C3
data8 0x8e39d9cd73464364 , 0x0000bffe // cos(11 pi/16)  -S3

data8 0xb504f333f9de6484 , 0x00003ffe // sin(12 pi/16)  S4
data8 0xb504f333f9de6484 , 0x0000bffe // cos(12 pi/16)  -S4


data8 0x8e39d9cd73464364 , 0x00003ffe // sin(13 pi/16) S3
data8 0xd4db3148750d181a , 0x0000bffe // cos(13 pi/16) -C3

data8 0xc3ef1535754b168e , 0x00003ffd // sin(14 pi/16) S2
data8 0xec835e79946a3146 , 0x0000bffe // cos(14 pi/16) -C2

data8 0xc7c5c1e34d3055b3 , 0x00003ffc // sin(15 pi/16) S1
data8 0xfb14be7fbae58157 , 0x0000bffe // cos(15 pi/16) -C1

data8 0x0000000000000000 , 0x00000000 // sin(16 pi/16) S0
data8 0x8000000000000000 , 0x0000bfff // cos(16 pi/16) -C0


data8 0xc7c5c1e34d3055b3 , 0x0000bffc // sin(17 pi/16) -S1
data8 0xfb14be7fbae58157 , 0x0000bffe // cos(17 pi/16) -C1

data8 0xc3ef1535754b168e , 0x0000bffd // sin(18 pi/16) -S2
data8 0xec835e79946a3146 , 0x0000bffe // cos(18 pi/16) -C2

data8 0x8e39d9cd73464364 , 0x0000bffe // sin(19 pi/16) -S3
data8 0xd4db3148750d181a , 0x0000bffe // cos(19 pi/16) -C3

data8 0xb504f333f9de6484 , 0x0000bffe // sin(20 pi/16) -S4
data8 0xb504f333f9de6484 , 0x0000bffe // cos(20 pi/16) -S4


data8 0xd4db3148750d181a , 0x0000bffe // sin(21 pi/16) -C3
data8 0x8e39d9cd73464364 , 0x0000bffe // cos(21 pi/16) -S3

data8 0xec835e79946a3146 , 0x0000bffe // sin(22 pi/16) -C2
data8 0xc3ef1535754b168e , 0x0000bffd // cos(22 pi/16) -S2

data8 0xfb14be7fbae58157 , 0x0000bffe // sin(23 pi/16) -C1
data8 0xc7c5c1e34d3055b3 , 0x0000bffc // cos(23 pi/16) -S1

data8 0x8000000000000000 , 0x0000bfff // sin(24 pi/16) -C0
data8 0x0000000000000000 , 0x00000000 // cos(24 pi/16) S0


data8 0xfb14be7fbae58157 , 0x0000bffe // sin(25 pi/16) -C1
data8 0xc7c5c1e34d3055b3 , 0x00003ffc // cos(25 pi/16) S1

data8 0xec835e79946a3146 , 0x0000bffe // sin(26 pi/16) -C2
data8 0xc3ef1535754b168e , 0x00003ffd // cos(26 pi/16) S2

data8 0xd4db3148750d181a , 0x0000bffe // sin(27 pi/16) -C3
data8 0x8e39d9cd73464364 , 0x00003ffe // cos(27 pi/16) S3

data8 0xb504f333f9de6484 , 0x0000bffe // sin(28 pi/16) -S4
data8 0xb504f333f9de6484 , 0x00003ffe // cos(28 pi/16) S4


data8 0x8e39d9cd73464364 , 0x0000bffe // sin(29 pi/16) -S3
data8 0xd4db3148750d181a , 0x00003ffe // cos(29 pi/16) C3

data8 0xc3ef1535754b168e , 0x0000bffd // sin(30 pi/16) -S2
data8 0xec835e79946a3146 , 0x00003ffe // cos(30 pi/16) C2

data8 0xc7c5c1e34d3055b3 , 0x0000bffc // sin(31 pi/16) -S1
data8 0xfb14be7fbae58157 , 0x00003ffe // cos(31 pi/16) C1

data8 0x0000000000000000 , 0x00000000 // sin(32 pi/16) S0
data8 0x8000000000000000 , 0x00003fff // cos(32 pi/16) C0
ASM_SIZE_DIRECTIVE(double_sin_cos_beta_k4)

.align 32
.global sin#
.global cos#
#ifdef _LIBC
.global __sin#
.global __cos#
#endif

////////////////////////////////////////////////////////
// There are two entry points: sin and cos


// If from sin, p8 is true
// If from cos, p9 is true

.section .text
.proc  sin#
#ifdef _LIBC
.proc  __sin#
#endif
.align 32

sin:
#ifdef _LIBC
__sin:
#endif

{ .mlx
      alloc          r32=ar.pfs,1,13,0,0
      movl sind_GR_sig_inv_pi_by_16 = 0xA2F9836E4E44152A // significand of 16/pi
}
{ .mlx
      addl           sind_AD_1   = @ltoff(double_sind_pi), gp
      movl sind_GR_rshf_2to61 = 0x47b8000000000000 // 1.1000 2^(63+63-2)
}
;;

{ .mfi
      ld8 sind_AD_1 = [sind_AD_1]
      fnorm     sind_NORM_f8  = f8
      cmp.eq     p8,p9         = r0, r0
}
{ .mib
      mov sind_GR_exp_2tom61 = 0xffff-61 // exponent of scaling factor 2^-61
      mov            sind_r_sincos = 0x0
      br.cond.sptk   L(SIND_SINCOS)
}
;;

.endp sin
ASM_SIZE_DIRECTIVE(sin)


.section .text
.proc  cos#
#ifdef _LIBC
.proc  __cos#
#endif
.align 32
cos:
#ifdef _LIBC
__cos:
#endif

{ .mlx
      alloc          r32=ar.pfs,1,13,0,0
      movl sind_GR_sig_inv_pi_by_16 = 0xA2F9836E4E44152A // significand of 16/pi
}
{ .mlx
      addl           sind_AD_1   = @ltoff(double_sind_pi), gp
      movl sind_GR_rshf_2to61 = 0x47b8000000000000 // 1.1000 2^(63+63-2)
}
;;

{ .mfi
      ld8 sind_AD_1 = [sind_AD_1]
      fnorm.s1     sind_NORM_f8  = f8
      cmp.eq     p9,p8         = r0, r0
}
{ .mib
      mov sind_GR_exp_2tom61 = 0xffff-61 // exponent of scaling factor 2^-61
      mov            sind_r_sincos = 0x8
      br.cond.sptk   L(SIND_SINCOS)
}
;;


////////////////////////////////////////////////////////
// All entry points end up here.
// If from sin, sind_r_sincos is 0 and p8 is true
// If from cos, sind_r_sincos is 8 = 2^(k-1) and p9 is true
// We add sind_r_sincos to N

L(SIND_SINCOS):


// Form two constants we need
//  16/pi * 2^-2 * 2^63, scaled by 2^61 since we just loaded the significand
//  1.1000...000 * 2^(63+63-2) to right shift int(W) into the low significand
// fcmp used to set denormal, and invalid on snans
{ .mfi
      setf.sig sind_SIG_INV_PI_BY_16_2TO61 = sind_GR_sig_inv_pi_by_16
      fcmp.eq.s0 p12,p0=f8,f0
      mov       sind_r_17_ones    = 0x1ffff
}
{ .mlx
      setf.d sind_RSHF_2TO61 = sind_GR_rshf_2to61
      movl sind_GR_rshf = 0x43e8000000000000 // 1.1000 2^63 for right shift
}
;;

// Form another constant
//  2^-61 for scaling Nfloat
// 0x10009 is register_bias + 10.
// So if f8 > 2^10 = Gamma, go to DBX
{ .mfi
      setf.exp sind_2TOM61 = sind_GR_exp_2tom61
      fclass.m  p13,p0 = f8, 0x23           // Test for x inf
      mov       sind_exp_limit = 0x10009
}
;;

// Load the two pieces of pi/16
// Form another constant
//  1.1000...000 * 2^63, the right shift constant
{ .mmf
      ldfe      sind_Pi_by_16_hi  = [sind_AD_1],16
      setf.d sind_RSHF = sind_GR_rshf
      fclass.m  p14,p0 = f8, 0xc3           // Test for x nan
}
;;

{ .mfi
      ldfe      sind_Pi_by_16_lo  = [sind_AD_1],16
(p13) frcpa.s0 f8,p12=f0,f0               // force qnan indef for x=inf
      addl gr_tmp = -1,r0
}
{ .mfb
      addl           sind_AD_beta_table   = @ltoff(double_sin_cos_beta_k4), gp
      nop.f 999
(p13) br.ret.spnt    b0 ;;                // Exit for x=inf
}

// Start loading P, Q coefficients
// SIN(0)
{ .mfi
      ldfpd      sind_P4,sind_Q4 = [sind_AD_1],16
(p8)  fclass.m.unc  p6,p0 = f8, 0x07      // Test for sin(0)
      nop.i 999
}
{ .mfb
      addl           sind_AD_beta_table   = @ltoff(double_sin_cos_beta_k4), gp
(p14) fma.d f8=f8,f1,f0                   // qnan for x=nan
(p14) br.ret.spnt    b0 ;;                // Exit for x=nan
}


// COS(0)
{ .mfi
      getf.exp  sind_r_signexp    = f8
(p9)  fclass.m.unc  p7,p0 = f8, 0x07      // Test for sin(0)
      nop.i 999
}
{ .mfi
      ld8 sind_AD_beta_table = [sind_AD_beta_table]
      nop.f 999
      nop.i 999 ;;
}

{ .mmb
      ldfpd      sind_P3,sind_Q3 = [sind_AD_1],16
      setf.sig fp_tmp = gr_tmp // Create constant such that fmpy sets inexact
(p6)  br.ret.spnt    b0 ;;
}

{ .mfb
      and       sind_r_exp = sind_r_17_ones, sind_r_signexp
(p7)  fmerge.s      f8 = f1,f1
(p7)  br.ret.spnt    b0 ;;
}

// p10 is true if we must call routines to handle larger arguments
// p10 is true if f8 exp is > 0x10009

{ .mfi
      ldfpd      sind_P2,sind_Q2 = [sind_AD_1],16
      nop.f 999
      cmp.ge  p10,p0 = sind_r_exp,sind_exp_limit
}
;;

// sind_W          = x * sind_Inv_Pi_by_16
// Multiply x by scaled 16/pi and add large const to shift integer part of W to
//   rightmost bits of significand
{ .mfi
      ldfpd      sind_P1,sind_Q1 = [sind_AD_1]
      fma.s1 sind_W_2TO61_RSH = sind_NORM_f8,sind_SIG_INV_PI_BY_16_2TO61,sind_RSHF_2TO61
      nop.i 999
}
{ .mbb
(p10) cmp.ne.unc p11,p12=sind_r_sincos,r0  // p11 call __libm_cos_double_dbx
                                           // p12 call __libm_sin_double_dbx
(p11) br.cond.spnt L(COSD_DBX)
(p12) br.cond.spnt L(SIND_DBX)
}
;;


// sind_NFLOAT = Round_Int_Nearest(sind_W)
// This is done by scaling back by 2^-61 and subtracting the shift constant
{ .mfi
      nop.m 999
      fms.s1 sind_NFLOAT = sind_W_2TO61_RSH,sind_2TOM61,sind_RSHF
      nop.i 999 ;;
}


// get N = (int)sind_int_Nfloat
{ .mfi
      getf.sig  sind_GR_n = sind_W_2TO61_RSH
      nop.f 999
      nop.i 999 ;;
}

// Add 2^(k-1) (which is in sind_r_sincos) to N
// sind_r          = -sind_Nfloat * sind_Pi_by_16_hi + x
// sind_r          = sind_r -sind_Nfloat * sind_Pi_by_16_lo
{ .mfi
      add       sind_GR_n = sind_GR_n, sind_r_sincos
      fnma.s1  sind_r      = sind_NFLOAT, sind_Pi_by_16_hi, sind_NORM_f8
      nop.i 999 ;;
}


// Get M (least k+1 bits of N)
{ .mmi
      and       sind_GR_m = 0x1f,sind_GR_n ;;
      nop.m 999
      shl       sind_GR_32m = sind_GR_m,5 ;;
}

// Add 32*M to address of sin_cos_beta table
{ .mmi
      add       sind_AD_2 = sind_GR_32m, sind_AD_beta_table
      nop.m 999
      nop.i 999 ;;
}

{ .mfi
      ldfe      sind_Sm = [sind_AD_2],16
(p8)  fclass.m.unc p10,p0=f8,0x0b  // If sin, note denormal input to set uflow
      nop.i 999 ;;
}

{ .mfi
      ldfe      sind_Cm = [sind_AD_2]
      fnma.s1  sind_r      = sind_NFLOAT, sind_Pi_by_16_lo,  sind_r
      nop.i 999 ;;
}

// get rsq
{ .mfi
      nop.m 999
      fma.s1   sind_rsq  = sind_r, sind_r,   f0
      nop.i 999
}
{ .mfi
      nop.m 999
      fmpy.s0  fp_tmp = fp_tmp,fp_tmp // fmpy forces inexact flag
      nop.i 999 ;;
}

// form P and Q series
{ .mfi
      nop.m 999
      fma.s1      sind_P_temp1 = sind_rsq, sind_P4, sind_P3
      nop.i 999
}

{ .mfi
      nop.m 999
      fma.s1      sind_Q_temp1 = sind_rsq, sind_Q4, sind_Q3
      nop.i 999 ;;
}

// get rcube and sm*rsq
{ .mfi
      nop.m 999
      fmpy.s1     sind_srsq    = sind_Sm,sind_rsq
      nop.i 999
}

{ .mfi
      nop.m 999
      fmpy.s1     sind_rcub    = sind_r, sind_rsq
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
      fma.s1      sind_Q_temp2 = sind_rsq, sind_Q_temp1, sind_Q2
      nop.i 999
}

{ .mfi
      nop.m 999
      fma.s1      sind_P_temp2 = sind_rsq, sind_P_temp1, sind_P2
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
      fma.s1      sind_Q       = sind_rsq, sind_Q_temp2, sind_Q1
      nop.i 999
}

{ .mfi
      nop.m 999
      fma.s1      sind_P       = sind_rsq, sind_P_temp2, sind_P1
      nop.i 999 ;;
}

// Get final P and Q
{ .mfi
      nop.m 999
      fma.s1   sind_Q = sind_srsq,sind_Q, sind_Sm
      nop.i 999
}

{ .mfi
      nop.m 999
      fma.s1   sind_P = sind_rcub,sind_P, sind_r
      nop.i 999 ;;
}

// If sin(denormal), force inexact to be set
{ .mfi
      nop.m 999
(p10) fmpy.d.s0 fp_tmp = f8,f8
      nop.i 999 ;;
}

// Final calculation
{ .mfb
      nop.m 999
      fma.d    f8     = sind_Cm, sind_P, sind_Q
      br.ret.sptk    b0 ;;
}
.endp cos#
ASM_SIZE_DIRECTIVE(cos#)



.proc __libm_callout_1s
__libm_callout_1s:
L(SIND_DBX):
.prologue
{ .mfi
        nop.m 0
        nop.f 0
.save   ar.pfs,GR_SAVE_PFS
        mov  GR_SAVE_PFS=ar.pfs
}
;;

{ .mfi
        mov GR_SAVE_GP=gp
        nop.f 0
.save   b0, GR_SAVE_B0
        mov GR_SAVE_B0=b0
}

.body
{ .mib
      nop.m 999
      nop.i 999
      br.call.sptk.many   b0=__libm_sin_double_dbx# ;;
}
;;


{ .mfi
       mov gp        = GR_SAVE_GP
       nop.f  999
       mov b0        = GR_SAVE_B0
}
;;

{ .mib
      nop.m 999
      mov ar.pfs    = GR_SAVE_PFS
      br.ret.sptk     b0 ;;
}
.endp  __libm_callout_1s
ASM_SIZE_DIRECTIVE(__libm_callout_1s)


.proc __libm_callout_1c
__libm_callout_1c:
L(COSD_DBX):
.prologue
{ .mfi
        nop.m 0
        nop.f 0
.save   ar.pfs,GR_SAVE_PFS
        mov  GR_SAVE_PFS=ar.pfs
}
;;

{ .mfi
        mov GR_SAVE_GP=gp
        nop.f 0
.save   b0, GR_SAVE_B0
        mov GR_SAVE_B0=b0
}

.body
{ .mib
      nop.m 999
      nop.i 999
      br.call.sptk.many   b0=__libm_cos_double_dbx# ;;
}
;;


{ .mfi
       mov gp        = GR_SAVE_GP
       nop.f  999
       mov b0        = GR_SAVE_B0
}
;;

{ .mib
      nop.m 999
      mov ar.pfs    = GR_SAVE_PFS
      br.ret.sptk     b0 ;;
}
.endp  __libm_callout_1c
ASM_SIZE_DIRECTIVE(__libm_callout_1c)


// ====================================================================
// ====================================================================

// These functions calculate the sin and cos for inputs
// greater than 2^10
// __libm_sin_double_dbx# and __libm_cos_double_dbx#

// *********************************************************************
// *********************************************************************
//
// Function:   Combined sin(x) and cos(x), where
//
//             sin(x) = sine(x), for double precision x values
//             cos(x) = cosine(x), for double precision x values
//
// *********************************************************************
//
// Accuracy:       Within .7 ulps for 80-bit floating point values
//                 Very accurate for double precision values
//
// *********************************************************************
//
// Resources Used:
//
//    Floating-Point Registers: f8 (Input and Return Value)
//                              f32-f99
//
//    General Purpose Registers:
//      r32-r43
//      r44-r45 (Used to pass arguments to pi_by_2 reduce routine)
//
//    Predicate Registers:      p6-p13
//
// *********************************************************************
//
//  IEEE Special Conditions:
//
//    Denormal  fault raised on denormal inputs
//    Overflow exceptions do not occur
//    Underflow exceptions raised when appropriate for sin
//    (No specialized error handling for this routine)
//    Inexact raised when appropriate by algorithm
//
//    sin(SNaN) = QNaN
//    sin(QNaN) = QNaN
//    sin(inf) = QNaN
//    sin(+/-0) = +/-0
//    cos(inf) = QNaN
//    cos(SNaN) = QNaN
//    cos(QNaN) = QNaN
//    cos(0) = 1
//
// *********************************************************************
//
//  Mathematical Description
//  ========================
//
//  The computation of FSIN and FCOS is best handled in one piece of
//  code. The main reason is that given any argument Arg, computation
//  of trigonometric functions first calculate N and an approximation
//  to alpha where
//
//  Arg = N pi/2 + alpha, |alpha| <= pi/4.
//
//  Since
//
//  cos( Arg ) = sin( (N+1) pi/2 + alpha ),
//
//  therefore, the code for computing sine will produce cosine as long
//  as 1 is added to N immediately after the argument reduction
//  process.
//
//  Let M = N if sine
//      N+1 if cosine.
//
//  Now, given
//
//  Arg = M pi/2  + alpha, |alpha| <= pi/4,
//
//  let I = M mod 4, or I be the two lsb of M when M is represented
//  as 2's complement. I = [i_0 i_1]. Then
//
//  sin( Arg ) = (-1)^i_0  sin( alpha )	if i_1 = 0,
//             = (-1)^i_0  cos( alpha )     if i_1 = 1.
//
//  For example:
//       if M = -1, I = 11
//         sin ((-pi/2 + alpha) = (-1) cos (alpha)
//       if M = 0, I = 00
//         sin (alpha) = sin (alpha)
//       if M = 1, I = 01
//         sin (pi/2 + alpha) = cos (alpha)
//       if M = 2, I = 10
//         sin (pi + alpha) = (-1) sin (alpha)
//       if M = 3, I = 11
//         sin ((3/2)pi + alpha) = (-1) cos (alpha)
//
//  The value of alpha is obtained by argument reduction and
//  represented by two working precision numbers r and c where
//
//  alpha =  r  +  c     accurately.
//
//  The reduction method is described in a previous write up.
//  The argument reduction scheme identifies 4 cases. For Cases 2
//  and 4, because |alpha| is small, sin(r+c) and cos(r+c) can be
//  computed very easily by 2 or 3 terms of the Taylor series
//  expansion as follows:
//
//  Case 2:
//  -------
//
//  sin(r + c) = r + c - r^3/6	accurately
//  cos(r + c) = 1 - 2^(-67)	accurately
//
//  Case 4:
//  -------
//
//  sin(r + c) = r + c - r^3/6 + r^5/120	accurately
//  cos(r + c) = 1 - r^2/2 + r^4/24		accurately
//
//  The only cases left are Cases 1 and 3 of the argument reduction
//  procedure. These two cases will be merged since after the
//  argument is reduced in either cases, we have the reduced argument
//  represented as r + c and that the magnitude |r + c| is not small
//  enough to allow the usage of a very short approximation.
//
//  The required calculation is either
//
//  sin(r + c)  =  sin(r)  +  correction,  or
//  cos(r + c)  =  cos(r)  +  correction.
//
//  Specifically,
//
//	sin(r + c) = sin(r) + c sin'(r) + O(c^2)
//		   = sin(r) + c cos (r) + O(c^2)
//		   = sin(r) + c(1 - r^2/2)  accurately.
//  Similarly,
//
//	cos(r + c) = cos(r) - c sin(r) + O(c^2)
//		   = cos(r) - c(r - r^3/6)  accurately.
//
//  We therefore concentrate on accurately calculating sin(r) and
//  cos(r) for a working-precision number r, |r| <= pi/4 to within
//  0.1% or so.
//
//  The greatest challenge of this task is that the second terms of
//  the Taylor series
//
//	r - r^3/3! + r^r/5! - ...
//
//  and
//
//	1 - r^2/2! + r^4/4! - ...
//
//  are not very small when |r| is close to pi/4 and the rounding
//  errors will be a concern if simple polynomial accumulation is
//  used. When |r| < 2^-3, however, the second terms will be small
//  enough (6 bits or so of right shift) that a normal Horner
//  recurrence suffices. Hence there are two cases that we consider
//  in the accurate computation of sin(r) and cos(r), |r| <= pi/4.
//
//  Case small_r: |r| < 2^(-3)
//  --------------------------
//
//  Since Arg = M pi/4 + r + c accurately, and M mod 4 is [i_0 i_1],
//  we have
//
//	sin(Arg) = (-1)^i_0 * sin(r + c)	if i_1 = 0
//		 = (-1)^i_0 * cos(r + c) 	if i_1 = 1
//
//  can be accurately approximated by
//
//  sin(Arg) = (-1)^i_0 * [sin(r) + c]	if i_1 = 0
//           = (-1)^i_0 * [cos(r) - c*r] if i_1 = 1
//
//  because |r| is small and thus the second terms in the correction
//  are unneccessary.
//
//  Finally, sin(r) and cos(r) are approximated by polynomials of
//  moderate lengths.
//
//  sin(r) =  r + S_1 r^3 + S_2 r^5 + ... + S_5 r^11
//  cos(r) =  1 + C_1 r^2 + C_2 r^4 + ... + C_5 r^10
//
//  We can make use of predicates to selectively calculate
//  sin(r) or cos(r) based on i_1.
//
//  Case normal_r: 2^(-3) <= |r| <= pi/4
//  ------------------------------------
//
//  This case is more likely than the previous one if one considers
//  r to be uniformly distributed in [-pi/4 pi/4]. Again,
//
//  sin(Arg) = (-1)^i_0 * sin(r + c)	if i_1 = 0
//           = (-1)^i_0 * cos(r + c) 	if i_1 = 1.
//
//  Because |r| is now larger, we need one extra term in the
//  correction. sin(Arg) can be accurately approximated by
//
//  sin(Arg) = (-1)^i_0 * [sin(r) + c(1-r^2/2)]      if i_1 = 0
//           = (-1)^i_0 * [cos(r) - c*r*(1 - r^2/6)]    i_1 = 1.
//
//  Finally, sin(r) and cos(r) are approximated by polynomials of
//  moderate lengths.
//
//	sin(r) =  r + PP_1_hi r^3 + PP_1_lo r^3 +
//	              PP_2 r^5 + ... + PP_8 r^17
//
//	cos(r) =  1 + QQ_1 r^2 + QQ_2 r^4 + ... + QQ_8 r^16
//
//  where PP_1_hi is only about 16 bits long and QQ_1 is -1/2.
//  The crux in accurate computation is to calculate
//
//  r + PP_1_hi r^3   or  1 + QQ_1 r^2
//
//  accurately as two pieces: U_hi and U_lo. The way to achieve this
//  is to obtain r_hi as a 10 sig. bit number that approximates r to
//  roughly 8 bits or so of accuracy. (One convenient way is
//
//  r_hi := frcpa( frcpa( r ) ).)
//
//  This way,
//
//	r + PP_1_hi r^3 =  r + PP_1_hi r_hi^3 +
//	                        PP_1_hi (r^3 - r_hi^3)
//		        =  [r + PP_1_hi r_hi^3]  +
//			   [PP_1_hi (r - r_hi)
//			      (r^2 + r_hi r + r_hi^2) ]
//		        =  U_hi  +  U_lo
//
//  Since r_hi is only 10 bit long and PP_1_hi is only 16 bit long,
//  PP_1_hi * r_hi^3 is only at most 46 bit long and thus computed
//  exactly. Furthermore, r and PP_1_hi r_hi^3 are of opposite sign
//  and that there is no more than 8 bit shift off between r and
//  PP_1_hi * r_hi^3. Hence the sum, U_hi, is representable and thus
//  calculated without any error. Finally, the fact that
//
//	|U_lo| <= 2^(-8) |U_hi|
//
//  says that U_hi + U_lo is approximating r + PP_1_hi r^3 to roughly
//  8 extra bits of accuracy.
//
//  Similarly,
//
//	1 + QQ_1 r^2  =  [1 + QQ_1 r_hi^2]  +
//	                    [QQ_1 (r - r_hi)(r + r_hi)]
//		      =  U_hi  +  U_lo.
//
//  Summarizing, we calculate r_hi = frcpa( frcpa( r ) ).
//
//  If i_1 = 0, then
//
//    U_hi := r + PP_1_hi * r_hi^3
//    U_lo := PP_1_hi * (r - r_hi) * (r^2 + r*r_hi + r_hi^2)
//    poly := PP_1_lo r^3 + PP_2 r^5 + ... + PP_8 r^17
//    correction := c * ( 1 + C_1 r^2 )
//
//  Else ...i_1 = 1
//
//    U_hi := 1 + QQ_1 * r_hi * r_hi
//    U_lo := QQ_1 * (r - r_hi) * (r + r_hi)
//    poly := QQ_2 * r^4 + QQ_3 * r^6 + ... + QQ_8 r^16
//    correction := -c * r * (1 + S_1 * r^2)
//
//  End
//
//  Finally,
//
//	V := poly + ( U_lo + correction )
//
//                 /    U_hi  +  V         if i_0 = 0
//	result := |
//                 \  (-U_hi) -  V         if i_0 = 1
//
//  It is important that in the last step, negation of U_hi is
//  performed prior to the subtraction which is to be performed in
//  the user-set rounding mode.
//
//
//  Algorithmic Description
//  =======================
//
//  The argument reduction algorithm is tightly integrated into FSIN
//  and FCOS which share the same code. The following is complete and
//  self-contained. The argument reduction description given
//  previously is repeated below.
//
//
//  Step 0. Initialization.
//
//   If FSIN is invoked, set N_inc := 0; else if FCOS is invoked,
//   set N_inc := 1.
//
//  Step 1. Check for exceptional and special cases.
//
//   * If Arg is +-0, +-inf, NaN, NaT, go to Step 10 for special
//     handling.
//   * If |Arg| < 2^24, go to Step 2 for reduction of moderate
//     arguments. This is the most likely case.
//   * If |Arg| < 2^63, go to Step 8 for pre-reduction of large
//     arguments.
//   * If |Arg| >= 2^63, go to Step 10 for special handling.
//
//  Step 2. Reduction of moderate arguments.
//
//  If |Arg| < pi/4 	...quick branch
//     N_fix := N_inc	(integer)
//     r     := Arg
//     c     := 0.0
//     Branch to Step 4, Case_1_complete
//  Else 		...cf. argument reduction
//     N     := Arg * two_by_PI	(fp)
//     N_fix := fcvt.fx( N )	(int)
//     N     := fcvt.xf( N_fix )
//     N_fix := N_fix + N_inc
//     s     := Arg - N * P_1	(first piece of pi/2)
//     w     := -N * P_2	(second piece of pi/2)
//
//     If |s| >= 2^(-33)
//        go to Step 3, Case_1_reduce
//     Else
//        go to Step 7, Case_2_reduce
//     Endif
//  Endif
//
//  Step 3. Case_1_reduce.
//
//  r := s + w
//  c := (s - r) + w	...observe order
//
//  Step 4. Case_1_complete
//
//  ...At this point, the reduced argument alpha is
//  ...accurately represented as r + c.
//  If |r| < 2^(-3), go to Step 6, small_r.
//
//  Step 5. Normal_r.
//
//  Let [i_0 i_1] by the 2 lsb of N_fix.
//  FR_rsq  := r * r
//  r_hi := frcpa( frcpa( r ) )
//  r_lo := r - r_hi
//
//  If i_1 = 0, then
//    poly := r*FR_rsq*(PP_1_lo + FR_rsq*(PP_2 + ... FR_rsq*PP_8))
//    U_hi := r + PP_1_hi*r_hi*r_hi*r_hi	...any order
//    U_lo := PP_1_hi*r_lo*(r*r + r*r_hi + r_hi*r_hi)
//    correction := c + c*C_1*FR_rsq		...any order
//  Else
//    poly := FR_rsq*FR_rsq*(QQ_2 + FR_rsq*(QQ_3 + ... + FR_rsq*QQ_8))
//    U_hi := 1 + QQ_1 * r_hi * r_hi		...any order
//    U_lo := QQ_1 * r_lo * (r + r_hi)
//    correction := -c*(r + S_1*FR_rsq*r)	...any order
//  Endif
//
//  V := poly + (U_lo + correction)	...observe order
//
//  result := (i_0 == 0?   1.0 : -1.0)
//
//  Last instruction in user-set rounding mode
//
//  result := (i_0 == 0?   result*U_hi + V :
//                        result*U_hi - V)
//
//  Return
//
//  Step 6. Small_r.
//
//  ...Use flush to zero mode without causing exception
//    Let [i_0 i_1] be the two lsb of N_fix.
//
//  FR_rsq := r * r
//
//  If i_1 = 0 then
//     z := FR_rsq*FR_rsq; z := FR_rsq*z *r
//     poly_lo := S_3 + FR_rsq*(S_4 + FR_rsq*S_5)
//     poly_hi := r*FR_rsq*(S_1 + FR_rsq*S_2)
//     correction := c
//     result := r
//  Else
//     z := FR_rsq*FR_rsq; z := FR_rsq*z
//     poly_lo := C_3 + FR_rsq*(C_4 + FR_rsq*C_5)
//     poly_hi := FR_rsq*(C_1 + FR_rsq*C_2)
//     correction := -c*r
//     result := 1
//  Endif
//
//  poly := poly_hi + (z * poly_lo + correction)
//
//  If i_0 = 1, result := -result
//
//  Last operation. Perform in user-set rounding mode
//
//  result := (i_0 == 0?     result + poly :
//                          result - poly )
//  Return
//
//  Step 7. Case_2_reduce.
//
//  ...Refer to the write up for argument reduction for
//  ...rationale. The reduction algorithm below is taken from
//  ...argument reduction description and integrated this.
//
//  w := N*P_3
//  U_1 := N*P_2 + w		...FMA
//  U_2 := (N*P_2 - U_1) + w	...2 FMA
//  ...U_1 + U_2 is  N*(P_2+P_3) accurately
//
//  r := s - U_1
//  c := ( (s - r) - U_1 ) - U_2
//
//  ...The mathematical sum r + c approximates the reduced
//  ...argument accurately. Note that although compared to
//  ...Case 1, this case requires much more work to reduce
//  ...the argument, the subsequent calculation needed for
//  ...any of the trigonometric function is very little because
//  ...|alpha| < 1.01*2^(-33) and thus two terms of the
//  ...Taylor series expansion suffices.
//
//  If i_1 = 0 then
//     poly := c + S_1 * r * r * r	...any order
//     result := r
//  Else
//     poly := -2^(-67)
//     result := 1.0
//  Endif
//
//  If i_0 = 1, result := -result
//
//  Last operation. Perform in user-set rounding mode
//
//  result := (i_0 == 0?     result + poly :
//                           result - poly )
//
//  Return
//
//
//  Step 8. Pre-reduction of large arguments.
//
//  ...Again, the following reduction procedure was described
//  ...in the separate write up for argument reduction, which
//  ...is tightly integrated here.

//  N_0 := Arg * Inv_P_0
//  N_0_fix := fcvt.fx( N_0 )
//  N_0 := fcvt.xf( N_0_fix)

//  Arg' := Arg - N_0 * P_0
//  w := N_0 * d_1
//  N := Arg' * two_by_PI
//  N_fix := fcvt.fx( N )
//  N := fcvt.xf( N_fix )
//  N_fix := N_fix + N_inc
//
//  s := Arg' - N * P_1
//  w := w - N * P_2
//
//  If |s| >= 2^(-14)
//     go to Step 3
//  Else
//     go to Step 9
//  Endif
//
//  Step 9. Case_4_reduce.
//
//    ...first obtain N_0*d_1 and -N*P_2 accurately
//   U_hi := N_0 * d_1		V_hi := -N*P_2
//   U_lo := N_0 * d_1 - U_hi	V_lo := -N*P_2 - U_hi	...FMAs
//
//   ...compute the contribution from N_0*d_1 and -N*P_3
//   w := -N*P_3
//   w := w + N_0*d_2
//   t := U_lo + V_lo + w		...any order
//
//   ...at this point, the mathematical value
//   ...s + U_hi + V_hi  + t approximates the true reduced argument
//   ...accurately. Just need to compute this accurately.
//
//   ...Calculate U_hi + V_hi accurately:
//   A := U_hi + V_hi
//   if |U_hi| >= |V_hi| then
//      a := (U_hi - A) + V_hi
//   else
//      a := (V_hi - A) + U_hi
//   endif
//   ...order in computing "a" must be observed. This branch is
//   ...best implemented by predicates.
//   ...A + a  is U_hi + V_hi accurately. Moreover, "a" is
//   ...much smaller than A: |a| <= (1/2)ulp(A).
//
//   ...Just need to calculate   s + A + a + t
//   C_hi := s + A		t := t + a
//   C_lo := (s - C_hi) + A
//   C_lo := C_lo + t
//
//   ...Final steps for reduction
//   r := C_hi + C_lo
//   c := (C_hi - r) + C_lo
//
//   ...At this point, we have r and c
//   ...And all we need is a couple of terms of the corresponding
//   ...Taylor series.
//
//   If i_1 = 0
//      poly := c + r*FR_rsq*(S_1 + FR_rsq*S_2)
//      result := r
//   Else
//      poly := FR_rsq*(C_1 + FR_rsq*C_2)
//      result := 1
//   Endif
//
//   If i_0 = 1, result := -result
//
//   Last operation. Perform in user-set rounding mode
//
//   result := (i_0 == 0?     result + poly :
//                            result - poly )
//   Return
//
//   Large Arguments: For arguments above 2**63, a Payne-Hanek
//   style argument reduction is used and pi_by_2 reduce is called.
//


#ifdef _LIBC
.rodata
#else
.data
#endif
.align 64

FSINCOS_CONSTANTS:
ASM_TYPE_DIRECTIVE(FSINCOS_CONSTANTS,@object)
data4 0x4B800000, 0xCB800000, 0x00000000,0x00000000 // two**24, -two**24
data4 0x4E44152A, 0xA2F9836E, 0x00003FFE,0x00000000 // Inv_pi_by_2
data4 0xCE81B9F1, 0xC84D32B0, 0x00004016,0x00000000 // P_0
data4 0x2168C235, 0xC90FDAA2, 0x00003FFF,0x00000000 // P_1
data4 0xFC8F8CBB, 0xECE675D1, 0x0000BFBD,0x00000000 // P_2
data4 0xACC19C60, 0xB7ED8FBB, 0x0000BF7C,0x00000000 // P_3
data4 0x5F000000, 0xDF000000, 0x00000000,0x00000000 // two_to_63, -two_to_63
data4 0x6EC6B45A, 0xA397E504, 0x00003FE7,0x00000000 // Inv_P_0
data4 0xDBD171A1, 0x8D848E89, 0x0000BFBF,0x00000000 // d_1
data4 0x18A66F8E, 0xD5394C36, 0x0000BF7C,0x00000000 // d_2
data4 0x2168C234, 0xC90FDAA2, 0x00003FFE,0x00000000 // pi_by_4
data4 0x2168C234, 0xC90FDAA2, 0x0000BFFE,0x00000000 // neg_pi_by_4
data4 0x3E000000, 0xBE000000, 0x00000000,0x00000000 // two**-3, -two**-3
data4 0x2F000000, 0xAF000000, 0x9E000000,0x00000000 // two**-33, -two**-33, -two**-67
data4 0xA21C0BC9, 0xCC8ABEBC, 0x00003FCE,0x00000000 // PP_8
data4 0x720221DA, 0xD7468A05, 0x0000BFD6,0x00000000 // PP_7
data4 0x640AD517, 0xB092382F, 0x00003FDE,0x00000000 // PP_6
data4 0xD1EB75A4, 0xD7322B47, 0x0000BFE5,0x00000000 // PP_5
data4 0xFFFFFFFE, 0xFFFFFFFF, 0x0000BFFD,0x00000000 // C_1
data4 0x00000000, 0xAAAA0000, 0x0000BFFC,0x00000000 // PP_1_hi
data4 0xBAF69EEA, 0xB8EF1D2A, 0x00003FEC,0x00000000 // PP_4
data4 0x0D03BB69, 0xD00D00D0, 0x0000BFF2,0x00000000 // PP_3
data4 0x88888962, 0x88888888, 0x00003FF8,0x00000000 // PP_2
data4 0xAAAB0000, 0xAAAAAAAA, 0x0000BFEC,0x00000000 // PP_1_lo
data4 0xC2B0FE52, 0xD56232EF, 0x00003FD2,0x00000000 // QQ_8
data4 0x2B48DCA6, 0xC9C99ABA, 0x0000BFDA,0x00000000 // QQ_7
data4 0x9C716658, 0x8F76C650, 0x00003FE2,0x00000000 // QQ_6
data4 0xFDA8D0FC, 0x93F27DBA, 0x0000BFE9,0x00000000 // QQ_5
data4 0xAAAAAAAA, 0xAAAAAAAA, 0x0000BFFC,0x00000000 // S_1
data4 0x00000000, 0x80000000, 0x0000BFFE,0x00000000 // QQ_1
data4 0x0C6E5041, 0xD00D00D0, 0x00003FEF,0x00000000 // QQ_4
data4 0x0B607F60, 0xB60B60B6, 0x0000BFF5,0x00000000 // QQ_3
data4 0xAAAAAA9B, 0xAAAAAAAA, 0x00003FFA,0x00000000 // QQ_2
data4 0xFFFFFFFE, 0xFFFFFFFF, 0x0000BFFD,0x00000000 // C_1
data4 0xAAAA719F, 0xAAAAAAAA, 0x00003FFA,0x00000000 // C_2
data4 0x0356F994, 0xB60B60B6, 0x0000BFF5,0x00000000 // C_3
data4 0xB2385EA9, 0xD00CFFD5, 0x00003FEF,0x00000000 // C_4
data4 0x292A14CD, 0x93E4BD18, 0x0000BFE9,0x00000000 // C_5
data4 0xAAAAAAAA, 0xAAAAAAAA, 0x0000BFFC,0x00000000 // S_1
data4 0x888868DB, 0x88888888, 0x00003FF8,0x00000000 // S_2
data4 0x055EFD4B, 0xD00D00D0, 0x0000BFF2,0x00000000 // S_3
data4 0x839730B9, 0xB8EF1C5D, 0x00003FEC,0x00000000 // S_4
data4 0xE5B3F492, 0xD71EA3A4, 0x0000BFE5,0x00000000 // S_5
data4 0x38800000, 0xB8800000, 0x00000000            // two**-14, -two**-14
ASM_SIZE_DIRECTIVE(FSINCOS_CONSTANTS)

FR_Input_X        = f8
FR_Neg_Two_to_M3  = f32
FR_Two_to_63      = f32
FR_Two_to_24      = f33
FR_Pi_by_4        = f33
FR_Two_to_M14     = f34
FR_Two_to_M33     = f35
FR_Neg_Two_to_24  = f36
FR_Neg_Pi_by_4    = f36
FR_Neg_Two_to_M14 = f37
FR_Neg_Two_to_M33 = f38
FR_Neg_Two_to_M67 = f39
FR_Inv_pi_by_2    = f40
FR_N_float        = f41
FR_N_fix          = f42
FR_P_1            = f43
FR_P_2            = f44
FR_P_3            = f45
FR_s              = f46
FR_w              = f47
FR_c              = f48
FR_r              = f49
FR_Z              = f50
FR_A              = f51
FR_a              = f52
FR_t              = f53
FR_U_1            = f54
FR_U_2            = f55
FR_C_1            = f56
FR_C_2            = f57
FR_C_3            = f58
FR_C_4            = f59
FR_C_5            = f60
FR_S_1            = f61
FR_S_2            = f62
FR_S_3            = f63
FR_S_4            = f64
FR_S_5            = f65
FR_poly_hi        = f66
FR_poly_lo        = f67
FR_r_hi           = f68
FR_r_lo           = f69
FR_rsq            = f70
FR_r_cubed        = f71
FR_C_hi           = f72
FR_N_0            = f73
FR_d_1            = f74
FR_V              = f75
FR_V_hi           = f75
FR_V_lo           = f76
FR_U_hi           = f77
FR_U_lo           = f78
FR_U_hiabs        = f79
FR_V_hiabs        = f80
FR_PP_8           = f81
FR_QQ_8           = f81
FR_PP_7           = f82
FR_QQ_7           = f82
FR_PP_6           = f83
FR_QQ_6           = f83
FR_PP_5           = f84
FR_QQ_5           = f84
FR_PP_4           = f85
FR_QQ_4           = f85
FR_PP_3           = f86
FR_QQ_3           = f86
FR_PP_2           = f87
FR_QQ_2           = f87
FR_QQ_1           = f88
FR_N_0_fix        = f89
FR_Inv_P_0        = f90
FR_corr           = f91
FR_poly           = f92
FR_d_2            = f93
FR_Two_to_M3      = f94
FR_Neg_Two_to_63  = f94
FR_P_0            = f95
FR_C_lo           = f96
FR_PP_1           = f97
FR_PP_1_lo        = f98
FR_ArgPrime       = f99

GR_Table_Base  = r32
GR_Table_Base1 = r33
GR_i_0         = r34
GR_i_1         = r35
GR_N_Inc       = r36
GR_Sin_or_Cos  = r37

GR_SAVE_B0     = r39
GR_SAVE_GP     = r40
GR_SAVE_PFS    = r41

.section .text
.proc __libm_sin_double_dbx#
.align 64
__libm_sin_double_dbx:

{ .mlx
alloc GR_Table_Base = ar.pfs,0,12,2,0
       movl GR_Sin_or_Cos = 0x0 ;;
}

{ .mmi
      nop.m 999
      addl           GR_Table_Base   = @ltoff(FSINCOS_CONSTANTS#), gp
      nop.i 999
}
;;

{ .mmi
      ld8 GR_Table_Base = [GR_Table_Base]
      nop.m 999
      nop.i 999
}
;;


{ .mib
      nop.m 999
      nop.i 999
       br.cond.sptk L(SINCOS_CONTINUE) ;;
}

.endp __libm_sin_double_dbx#
ASM_SIZE_DIRECTIVE(__libm_sin_double_dbx)

.section .text
.proc __libm_cos_double_dbx#
__libm_cos_double_dbx:

{ .mlx
alloc GR_Table_Base= ar.pfs,0,12,2,0
       movl GR_Sin_or_Cos = 0x1 ;;
}

{ .mmi
      nop.m 999
      addl           GR_Table_Base   = @ltoff(FSINCOS_CONSTANTS#), gp
      nop.i 999
}
;;

{ .mmi
      ld8 GR_Table_Base = [GR_Table_Base]
      nop.m 999
      nop.i 999
}
;;

//
//     Load Table Address
//
L(SINCOS_CONTINUE):

{ .mmi
       add GR_Table_Base1 = 96, GR_Table_Base
       ldfs	FR_Two_to_24 = [GR_Table_Base], 4
       nop.i 999
}
;;

{ .mmi
      nop.m 999
//
//     Load 2**24, load 2**63.
//
       ldfs	FR_Neg_Two_to_24 = [GR_Table_Base], 12
       mov   r41 = ar.pfs ;;
}

{ .mfi
       ldfs	FR_Two_to_63 = [GR_Table_Base1], 4
//
//     Check for unnormals - unsupported operands. We do not want
//     to generate denormal exception
//     Check for NatVals, QNaNs, SNaNs, +/-Infs
//     Check for EM unsupporteds
//     Check for Zero
//
       fclass.m.unc  p6, p8 =  FR_Input_X, 0x1E3
       mov   r40 = gp ;;
}

{ .mfi
      nop.m 999
       fclass.nm.unc p8, p0 =  FR_Input_X, 0x1FF
// GR_Sin_or_Cos denotes
       mov   r39 = b0
}

{ .mfb
       ldfs	FR_Neg_Two_to_63 = [GR_Table_Base1], 12
       fclass.m.unc p10, p0 = FR_Input_X, 0x007
(p6)   br.cond.spnt L(SINCOS_SPECIAL) ;;
}

{ .mib
      nop.m 999
      nop.i 999
(p8)   br.cond.spnt L(SINCOS_SPECIAL) ;;
}

{ .mib
      nop.m 999
      nop.i 999
//
//     Branch if +/- NaN, Inf.
//     Load -2**24, load -2**63.
//
(p10)  br.cond.spnt L(SINCOS_ZERO) ;;
}

{ .mmb
       ldfe	FR_Inv_pi_by_2 = [GR_Table_Base], 16
       ldfe	FR_Inv_P_0 = [GR_Table_Base1], 16
      nop.b 999 ;;
}

{ .mmb
      nop.m 999
       ldfe		FR_d_1 = [GR_Table_Base1], 16
      nop.b 999 ;;
}
//
//     Raise possible denormal operand flag with useful fcmp
//     Is x <= -2**63
//     Load Inv_P_0 for pre-reduction
//     Load Inv_pi_by_2
//

{ .mmb
       ldfe		FR_P_0 = [GR_Table_Base], 16
       ldfe	FR_d_2 = [GR_Table_Base1], 16
      nop.b 999 ;;
}
//
//     Load P_0
//     Load d_1
//     Is x >= 2**63
//     Is x <= -2**24?
//

{ .mmi
       ldfe	FR_P_1 = [GR_Table_Base], 16 ;;
//
//     Load P_1
//     Load d_2
//     Is x >= 2**24?
//
       ldfe	FR_P_2 = [GR_Table_Base], 16
      nop.i 999 ;;
}

{ .mmf
      nop.m 999
       ldfe	FR_P_3 = [GR_Table_Base], 16
       fcmp.le.unc.s1	p7, p8 = FR_Input_X, FR_Neg_Two_to_24
}

{ .mfi
      nop.m 999
//
//     Branch if +/- zero.
//     Decide about the paths to take:
//     If -2**24 < FR_Input_X < 2**24 - CASE 1 OR 2
//     OTHERWISE - CASE 3 OR 4
//
       fcmp.le.unc.s0	p10, p11 = FR_Input_X, FR_Neg_Two_to_63
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p8)   fcmp.ge.s1 p7, p0 = FR_Input_X, FR_Two_to_24
      nop.i 999
}

{ .mfi
       ldfe	FR_Pi_by_4 = [GR_Table_Base1], 16
(p11)  fcmp.ge.s1	p10, p0 = FR_Input_X, FR_Two_to_63
      nop.i 999 ;;
}

{ .mmi
       ldfe	FR_Neg_Pi_by_4 = [GR_Table_Base1], 16 ;;
       ldfs	FR_Two_to_M3 = [GR_Table_Base1], 4
      nop.i 999 ;;
}

{ .mib
       ldfs	FR_Neg_Two_to_M3 = [GR_Table_Base1], 12
      nop.i 999
//
//     Load P_2
//     Load P_3
//     Load pi_by_4
//     Load neg_pi_by_4
//     Load 2**(-3)
//     Load -2**(-3).
//
(p10)  br.cond.spnt L(SINCOS_ARG_TOO_LARGE) ;;
}

{ .mib
      nop.m 999
      nop.i 999
//
//     Branch out if x >= 2**63. Use Payne-Hanek Reduction
//
(p7)   br.cond.spnt L(SINCOS_LARGER_ARG) ;;
}

{ .mfi
      nop.m 999
//
//     Branch if Arg <= -2**24 or Arg >= 2**24 and use pre-reduction.
//
       fma.s1	FR_N_float = FR_Input_X, FR_Inv_pi_by_2, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
       fcmp.lt.unc.s1	p6, p7 = FR_Input_X, FR_Pi_by_4
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     Select the case when |Arg| < pi/4
//     Else Select the case when |Arg| >= pi/4
//
       fcvt.fx.s1 FR_N_fix = FR_N_float
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     N  = Arg * 2/pi
//     Check if Arg < pi/4
//
(p6)   fcmp.gt.s1 p6, p7 = FR_Input_X, FR_Neg_Pi_by_4
      nop.i 999 ;;
}
//
//     Case 2: Convert integer N_fix back to normalized floating-point value.
//     Case 1: p8 is only affected  when p6 is set
//

{ .mfi
(p7)   ldfs FR_Two_to_M33 = [GR_Table_Base1], 4
//
//     Grab the integer part of N and call it N_fix
//
(p6)   fmerge.se FR_r = FR_Input_X, FR_Input_X
//     If |x| < pi/4, r = x and c = 0
//     lf |x| < pi/4, is x < 2**(-3).
//     r = Arg
//     c = 0
(p6)   mov GR_N_Inc = GR_Sin_or_Cos ;;
}

{ .mmf
      nop.m 999
(p7)   ldfs FR_Neg_Two_to_M33 = [GR_Table_Base1], 4
(p6)   fmerge.se FR_c = f0, f0
}

{ .mfi
      nop.m 999
(p6)   fcmp.lt.unc.s1	p8, p9 = FR_Input_X, FR_Two_to_M3
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     lf |x| < pi/4, is -2**(-3)< x < 2**(-3) - set p8.
//     If |x| >= pi/4,
//     Create the right N for |x| < pi/4 and otherwise
//     Case 2: Place integer part of N in GP register
//
(p7)   fcvt.xf FR_N_float = FR_N_fix
      nop.i 999 ;;
}

{ .mmf
      nop.m 999
(p7)   getf.sig	GR_N_Inc = FR_N_fix
(p8)   fcmp.gt.s1 p8, p0 = FR_Input_X, FR_Neg_Two_to_M3 ;;
}

{ .mib
      nop.m 999
      nop.i 999
//
//     Load 2**(-33), -2**(-33)
//
(p8)   br.cond.spnt L(SINCOS_SMALL_R) ;;
}

{ .mib
      nop.m 999
      nop.i 999
(p6)   br.cond.sptk L(SINCOS_NORMAL_R) ;;
}
//
//     if |x| < pi/4, branch based on |x| < 2**(-3) or otherwise.
//
//
//     In this branch, |x| >= pi/4.
//

{ .mfi
       ldfs FR_Neg_Two_to_M67 = [GR_Table_Base1], 8
//
//     Load -2**(-67)
//
       fnma.s1	FR_s = FR_N_float, FR_P_1, FR_Input_X
//
//     w = N * P_2
//     s = -N * P_1  + Arg
//
       add GR_N_Inc = GR_N_Inc, GR_Sin_or_Cos
}

{ .mfi
      nop.m 999
       fma.s1	FR_w = FR_N_float, FR_P_2, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     Adjust N_fix by N_inc to determine whether sine or
//     cosine is being calculated
//
       fcmp.lt.unc.s1 p7, p6 = FR_s, FR_Two_to_M33
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p7)   fcmp.gt.s1 p7, p6 = FR_s, FR_Neg_Two_to_M33
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//     Remember x >= pi/4.
//     Is s <= -2**(-33) or s >= 2**(-33) (p6)
//     or -2**(-33) < s < 2**(-33) (p7)
(p6)   fms.s1 FR_r = FR_s, f1, FR_w
      nop.i 999
}

{ .mfi
      nop.m 999
(p7)   fma.s1 FR_w = FR_N_float, FR_P_3, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p7)   fma.s1 FR_U_1 = FR_N_float, FR_P_2, FR_w
      nop.i 999
}

{ .mfi
      nop.m 999
(p6)   fms.s1 FR_c = FR_s, f1, FR_r
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     For big s: r = s - w: No futher reduction is necessary
//     For small s: w = N * P_3 (change sign) More reduction
//
(p6)   fcmp.lt.unc.s1 p8, p9 = FR_r, FR_Two_to_M3
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p8)   fcmp.gt.s1 p8, p9 = FR_r, FR_Neg_Two_to_M3
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p7)   fms.s1 FR_r = FR_s, f1, FR_U_1
      nop.i 999
}

{ .mfb
      nop.m 999
//
//     For big s: Is |r| < 2**(-3)?
//     For big s: c = S - r
//     For small s: U_1 = N * P_2 + w
//
//     If p8 is set, prepare to branch to Small_R.
//     If p9 is set, prepare to branch to Normal_R.
//     For big s,  r is complete here.
//
(p6)   fms.s1 FR_c = FR_c, f1, FR_w
//
//     For big s: c = c + w (w has not been negated.)
//     For small s: r = S - U_1
//
(p8)   br.cond.spnt	L(SINCOS_SMALL_R) ;;
}

{ .mib
      nop.m 999
      nop.i 999
(p9)   br.cond.sptk	L(SINCOS_NORMAL_R) ;;
}

{ .mfi
(p7)   add GR_Table_Base1 = 224, GR_Table_Base1
//
//     Branch to SINCOS_SMALL_R or SINCOS_NORMAL_R
//
(p7)   fms.s1 FR_U_2 = FR_N_float, FR_P_2, FR_U_1
//
//     c = S - U_1
//     r = S_1 * r
//
//
(p7)   extr.u	GR_i_1 = GR_N_Inc, 0, 1
}

{ .mmi
      nop.m 999 ;;
//
//     Get [i_0,i_1] - two lsb of N_fix_gr.
//     Do dummy fmpy so inexact is always set.
//
(p7)   cmp.eq.unc p9, p10 = 0x0, GR_i_1
(p7)   extr.u	GR_i_0 = GR_N_Inc, 1, 1 ;;
}
//
//     For small s: U_2 = N * P_2 - U_1
//     S_1 stored constant - grab the one stored with the
//     coefficients.
//

{ .mfi
(p7)   ldfe FR_S_1 = [GR_Table_Base1], 16
//
//     Check if i_1 and i_0  != 0
//
(p10)  fma.s1	FR_poly = f0, f1, FR_Neg_Two_to_M67
(p7)   cmp.eq.unc p11, p12 = 0x0, GR_i_0 ;;
}

{ .mfi
      nop.m 999
(p7)   fms.s1	FR_s = FR_s, f1, FR_r
      nop.i 999
}

{ .mfi
      nop.m 999
//
//     S = S - r
//     U_2 = U_2 + w
//     load S_1
//
(p7)   fma.s1	FR_rsq = FR_r, FR_r, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p7)   fma.s1	FR_U_2 = FR_U_2, f1, FR_w
      nop.i 999
}

{ .mfi
      nop.m 999
(p7)   fmerge.se FR_Input_X = FR_r, FR_r
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p10)  fma.s1 FR_Input_X = f0, f1, f1
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     FR_rsq = r * r
//     Save r as the result.
//
(p7)   fms.s1	FR_c = FR_s, f1, FR_U_1
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     if ( i_1 ==0) poly = c + S_1*r*r*r
//     else Result = 1
//
(p12)  fnma.s1 FR_Input_X = FR_Input_X, f1, f0
      nop.i 999
}

{ .mfi
      nop.m 999
(p7)   fma.s1	FR_r = FR_S_1, FR_r, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p7)   fma.d.s0	FR_S_1 = FR_S_1, FR_S_1, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     If i_1 != 0, poly = 2**(-67)
//
(p7)   fms.s1 FR_c = FR_c, f1, FR_U_2
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     c = c - U_2
//
(p9)   fma.s1 FR_poly = FR_r, FR_rsq, FR_c
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     i_0 != 0, so Result = -Result
//
(p11)  fma.d.s0 FR_Input_X = FR_Input_X, f1, FR_poly
      nop.i 999 ;;
}

{ .mfb
      nop.m 999
(p12)  fms.d.s0 FR_Input_X = FR_Input_X, f1, FR_poly
//
//     if (i_0 == 0),  Result = Result + poly
//     else            Result = Result - poly
//
       br.ret.sptk   b0 ;;
}
L(SINCOS_LARGER_ARG):

{ .mfi
      nop.m 999
       fma.s1 FR_N_0 = FR_Input_X, FR_Inv_P_0, f0
      nop.i 999
}
;;

//     This path for argument > 2*24
//     Adjust table_ptr1 to beginning of table.
//

{ .mmi
      nop.m 999
      addl           GR_Table_Base   = @ltoff(FSINCOS_CONSTANTS#), gp
      nop.i 999
}
;;

{ .mmi
      ld8 GR_Table_Base = [GR_Table_Base]
      nop.m 999
      nop.i 999
}
;;


//
//     Point to  2*-14
//     N_0 = Arg * Inv_P_0
//

{ .mmi
       add GR_Table_Base = 688, GR_Table_Base ;;
       ldfs FR_Two_to_M14 = [GR_Table_Base], 4
      nop.i 999 ;;
}

{ .mfi
       ldfs FR_Neg_Two_to_M14 = [GR_Table_Base], 0
      nop.f 999
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     Load values 2**(-14) and -2**(-14)
//
       fcvt.fx.s1 FR_N_0_fix = FR_N_0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     N_0_fix  = integer part of N_0
//
       fcvt.xf FR_N_0 = FR_N_0_fix
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     Make N_0 the integer part
//
       fnma.s1 FR_ArgPrime = FR_N_0, FR_P_0, FR_Input_X
      nop.i 999
}

{ .mfi
      nop.m 999
       fma.s1 FR_w = FR_N_0, FR_d_1, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     Arg' = -N_0 * P_0 + Arg
//     w  = N_0 * d_1
//
       fma.s1 FR_N_float = FR_ArgPrime, FR_Inv_pi_by_2, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     N = A' * 2/pi
//
       fcvt.fx.s1 FR_N_fix = FR_N_float
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     N_fix is the integer part
//
       fcvt.xf FR_N_float = FR_N_fix
      nop.i 999 ;;
}

{ .mfi
       getf.sig GR_N_Inc = FR_N_fix
      nop.f 999
      nop.i 999 ;;
}

{ .mii
      nop.m 999
      nop.i 999 ;;
       add GR_N_Inc = GR_N_Inc, GR_Sin_or_Cos ;;
}

{ .mfi
      nop.m 999
//
//     N is the integer part of the reduced-reduced argument.
//     Put the integer in a GP register
//
       fnma.s1 FR_s = FR_N_float, FR_P_1, FR_ArgPrime
      nop.i 999
}

{ .mfi
      nop.m 999
       fnma.s1 FR_w = FR_N_float, FR_P_2, FR_w
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     s = -N*P_1 + Arg'
//     w = -N*P_2 + w
//     N_fix_gr = N_fix_gr + N_inc
//
       fcmp.lt.unc.s1 p9, p8 = FR_s, FR_Two_to_M14
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p9)   fcmp.gt.s1 p9, p8 = FR_s, FR_Neg_Two_to_M14
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     For |s|  > 2**(-14) r = S + w (r complete)
//     Else       U_hi = N_0 * d_1
//
(p9)   fma.s1 FR_V_hi = FR_N_float, FR_P_2, f0
      nop.i 999
}

{ .mfi
      nop.m 999
(p9)   fma.s1 FR_U_hi = FR_N_0, FR_d_1, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     Either S <= -2**(-14) or S >= 2**(-14)
//     or -2**(-14) < s < 2**(-14)
//
(p8)   fma.s1 FR_r = FR_s, f1, FR_w
      nop.i 999
}

{ .mfi
      nop.m 999
(p9)   fma.s1 FR_w = FR_N_float, FR_P_3, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     We need abs of both U_hi and V_hi - don't
//     worry about switched sign of V_hi.
//
(p9)   fms.s1 FR_A = FR_U_hi, f1, FR_V_hi
      nop.i 999
}

{ .mfi
      nop.m 999
//
//     Big s: finish up c = (S - r) + w (c complete)
//     Case 4: A =  U_hi + V_hi
//     Note: Worry about switched sign of V_hi, so subtract instead of add.
//
(p9)   fnma.s1 FR_V_lo = FR_N_float, FR_P_2, FR_V_hi
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p9)   fms.s1 FR_U_lo = FR_N_0, FR_d_1, FR_U_hi
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p9)   fmerge.s FR_V_hiabs = f0, FR_V_hi
      nop.i 999
}

{ .mfi
      nop.m 999
//     For big s: c = S - r
//     For small s do more work: U_lo = N_0 * d_1 - U_hi
//
(p9)   fmerge.s FR_U_hiabs = f0, FR_U_hi
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     For big s: Is |r| < 2**(-3)
//     For big s: if p12 set, prepare to branch to Small_R.
//     For big s: If p13 set, prepare to branch to Normal_R.
//
(p8)   fms.s1 FR_c = FR_s, f1, FR_r
      nop.i 999
}

{ .mfi
      nop.m 999
//
//     For small S: V_hi = N * P_2
//                  w = N * P_3
//     Note the product does not include the (-) as in the writeup
//     so (-) missing for V_hi and w.
//
(p8)   fcmp.lt.unc.s1 p12, p13 = FR_r, FR_Two_to_M3
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p12)  fcmp.gt.s1 p12, p13 = FR_r, FR_Neg_Two_to_M3
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p8)   fma.s1 FR_c = FR_c, f1, FR_w
      nop.i 999
}

{ .mfb
      nop.m 999
(p9)   fms.s1 FR_w = FR_N_0, FR_d_2, FR_w
(p12)  br.cond.spnt L(SINCOS_SMALL_R) ;;
}

{ .mib
      nop.m 999
      nop.i 999
(p13)  br.cond.sptk L(SINCOS_NORMAL_R) ;;
}

{ .mfi
      nop.m 999
//
//     Big s: Vector off when |r| < 2**(-3).  Recall that p8 will be true.
//     The remaining stuff is for Case 4.
//     Small s: V_lo = N * P_2 + U_hi (U_hi is in place of V_hi in writeup)
//     Note: the (-) is still missing for V_lo.
//     Small s: w = w + N_0 * d_2
//     Note: the (-) is now incorporated in w.
//
(p9)   fcmp.ge.unc.s1 p10, p11 = FR_U_hiabs, FR_V_hiabs
       extr.u	GR_i_1 = GR_N_Inc, 0, 1 ;;
}

{ .mfi
      nop.m 999
//
//     C_hi = S + A
//
(p9)   fma.s1 FR_t = FR_U_lo, f1, FR_V_lo
       extr.u	GR_i_0 = GR_N_Inc, 1, 1 ;;
}

{ .mfi
      nop.m 999
//
//     t = U_lo + V_lo
//
//
(p10)  fms.s1 FR_a = FR_U_hi, f1, FR_A
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p11)  fma.s1 FR_a = FR_V_hi, f1, FR_A
      nop.i 999
}
;;

{ .mmi
      nop.m 999
      addl           GR_Table_Base   = @ltoff(FSINCOS_CONSTANTS#), gp
      nop.i 999
}
;;

{ .mmi
      ld8 GR_Table_Base = [GR_Table_Base]
      nop.m 999
      nop.i 999
}
;;


{ .mfi
       add GR_Table_Base = 528, GR_Table_Base
//
//     Is U_hiabs >= V_hiabs?
//
(p9)   fma.s1 FR_C_hi = FR_s, f1, FR_A
      nop.i 999 ;;
}

{ .mmi
       ldfe FR_C_1 = [GR_Table_Base], 16 ;;
       ldfe FR_C_2 = [GR_Table_Base], 64
      nop.i 999 ;;
}

{ .mmf
      nop.m 999
//
//     c = c + C_lo  finished.
//     Load  C_2
//
       ldfe	FR_S_1 = [GR_Table_Base], 16
//
//     C_lo = S - C_hi
//
       fma.s1 FR_t = FR_t, f1, FR_w ;;
}
//
//     r and c have been computed.
//     Make sure ftz mode is set - should be automatic when using wre
//     |r| < 2**(-3)
//     Get [i_0,i_1] - two lsb of N_fix.
//     Load S_1
//

{ .mfi
       ldfe FR_S_2 = [GR_Table_Base], 64
//
//     t = t + w
//
(p10)  fms.s1 FR_a = FR_a, f1, FR_V_hi
       cmp.eq.unc p9, p10 = 0x0, GR_i_0
}

{ .mfi
      nop.m 999
//
//     For larger u than v: a = U_hi - A
//     Else a = V_hi - A (do an add to account for missing (-) on V_hi
//
       fms.s1 FR_C_lo = FR_s, f1, FR_C_hi
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p11)  fms.s1 FR_a = FR_U_hi, f1, FR_a
       cmp.eq.unc p11, p12 = 0x0, GR_i_1
}

{ .mfi
      nop.m 999
//
//     If u > v: a = (U_hi - A)  + V_hi
//     Else      a = (V_hi - A)  + U_hi
//     In each case account for negative missing from V_hi.
//
       fma.s1 FR_C_lo = FR_C_lo, f1, FR_A
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     C_lo = (S - C_hi) + A
//
       fma.s1 FR_t = FR_t, f1, FR_a
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     t = t + a
//
       fma.s1 FR_C_lo = FR_C_lo, f1, FR_t
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     C_lo = C_lo + t
//     Adjust Table_Base to beginning of table
//
       fma.s1 FR_r = FR_C_hi, f1, FR_C_lo
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     Load S_2
//
       fma.s1 FR_rsq = FR_r, FR_r, f0
      nop.i 999
}

{ .mfi
      nop.m 999
//
//     Table_Base points to C_1
//     r = C_hi + C_lo
//
       fms.s1 FR_c = FR_C_hi, f1, FR_r
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     if i_1 ==0: poly = S_2 * FR_rsq + S_1
//     else        poly = C_2 * FR_rsq + C_1
//
(p11)  fma.s1 FR_Input_X = f0, f1, FR_r
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p12)  fma.s1 FR_Input_X = f0, f1, f1
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     Compute r_cube = FR_rsq * r
//
(p11)  fma.s1 FR_poly = FR_rsq, FR_S_2, FR_S_1
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p12)  fma.s1 FR_poly = FR_rsq, FR_C_2, FR_C_1
      nop.i 999
}

{ .mfi
      nop.m 999
//
//     Compute FR_rsq = r * r
//     Is i_1 == 0 ?
//
       fma.s1 FR_r_cubed = FR_rsq, FR_r, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     c = C_hi - r
//     Load  C_1
//
       fma.s1 FR_c = FR_c, f1, FR_C_lo
      nop.i 999
}

{ .mfi
      nop.m 999
//
//     if i_1 ==0: poly = r_cube * poly + c
//     else        poly = FR_rsq * poly
//
(p10)  fms.s1 FR_Input_X = f0, f1, FR_Input_X
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     if i_1 ==0: Result = r
//     else        Result = 1.0
//
(p11)  fma.s1 FR_poly = FR_r_cubed, FR_poly, FR_c
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p12)  fma.s1 FR_poly = FR_rsq, FR_poly, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//     if i_0 !=0: Result = -Result
//
(p9)   fma.d.s0 FR_Input_X = FR_Input_X, f1, FR_poly
      nop.i 999 ;;
}

{ .mfb
      nop.m 999
(p10)  fms.d.s0 FR_Input_X = FR_Input_X, f1, FR_poly
//
//     if i_0 == 0: Result = Result + poly
//     else         Result = Result - poly
//
       br.ret.sptk   b0 ;;
}
L(SINCOS_SMALL_R):

{ .mii
      nop.m 999
      	extr.u	GR_i_1 = GR_N_Inc, 0, 1 ;;
//
//
//      Compare both i_1 and i_0 with 0.
//      if i_1 == 0, set p9.
//      if i_0 == 0, set p11.
//
      	cmp.eq.unc p9, p10 = 0x0, GR_i_1 ;;
}

{ .mfi
      nop.m 999
      	fma.s1 FR_rsq = FR_r, FR_r, f0
      	extr.u	GR_i_0 = GR_N_Inc, 1, 1 ;;
}

{ .mfi
      nop.m 999
//
// 	Z = Z * FR_rsq
//
(p10)	fnma.s1	FR_c = FR_c, FR_r, f0
      	cmp.eq.unc p11, p12 = 0x0, GR_i_0
}
;;

// ******************************************************************
// ******************************************************************
// ******************************************************************
//      r and c have been computed.
//      We know whether this is the sine or cosine routine.
//      Make sure ftz mode is set - should be automatic when using wre
//      |r| < 2**(-3)
//
//      Set table_ptr1 to beginning of constant table.
//      Get [i_0,i_1] - two lsb of N_fix_gr.
//

{ .mmi
      nop.m 999
      addl           GR_Table_Base   = @ltoff(FSINCOS_CONSTANTS#), gp
      nop.i 999
}
;;

{ .mmi
      ld8 GR_Table_Base = [GR_Table_Base]
      nop.m 999
      nop.i 999
}
;;


//
//      Set table_ptr1 to point to S_5.
//      Set table_ptr1 to point to C_5.
//      Compute FR_rsq = r * r
//

{ .mfi
(p9)  	add GR_Table_Base = 672, GR_Table_Base
(p10)	fmerge.s FR_r = f1, f1
(p10) 	add GR_Table_Base = 592, GR_Table_Base ;;
}
//
//      Set table_ptr1 to point to S_5.
//      Set table_ptr1 to point to C_5.
//

{ .mmi
(p9)  	ldfe FR_S_5 = [GR_Table_Base], -16 ;;
//
//      if (i_1 == 0) load S_5
//      if (i_1 != 0) load C_5
//
(p9)  	ldfe FR_S_4 = [GR_Table_Base], -16
      nop.i 999 ;;
}

{ .mmf
(p10) 	ldfe FR_C_5 = [GR_Table_Base], -16
//
//      Z = FR_rsq * FR_rsq
//
(p9)  	ldfe FR_S_3 = [GR_Table_Base], -16
//
//      Compute FR_rsq = r * r
//      if (i_1 == 0) load S_4
//      if (i_1 != 0) load C_4
//
       	fma.s1 FR_Z = FR_rsq, FR_rsq, f0 ;;
}
//
//      if (i_1 == 0) load S_3
//      if (i_1 != 0) load C_3
//

{ .mmi
(p9)  	ldfe FR_S_2 = [GR_Table_Base], -16 ;;
//
//      if (i_1 == 0) load S_2
//      if (i_1 != 0) load C_2
//
(p9)  	ldfe FR_S_1 = [GR_Table_Base], -16
      nop.i 999
}

{ .mmi
(p10) 	ldfe FR_C_4 = [GR_Table_Base], -16 ;;
(p10)  	ldfe FR_C_3 = [GR_Table_Base], -16
      nop.i 999 ;;
}

{ .mmi
(p10) 	ldfe FR_C_2 = [GR_Table_Base], -16 ;;
(p10) 	ldfe FR_C_1 = [GR_Table_Base], -16
      nop.i 999
}

{ .mfi
      nop.m 999
//
//      if (i_1 != 0):
//      poly_lo = FR_rsq * C_5 + C_4
//      poly_hi = FR_rsq * C_2 + C_1
//
(p9)  	fma.s1 FR_Z = FR_Z, FR_r, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1 == 0) load S_1
//      if (i_1 != 0) load C_1
//
(p9)  	fma.s1 FR_poly_lo = FR_rsq, FR_S_5, FR_S_4
      nop.i 999
}

{ .mfi
      nop.m 999
//
//      c = -c * r
//      dummy fmpy's to flag inexact.
//
(p9)	fma.d.s0 FR_S_4 = FR_S_4, FR_S_4, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      poly_lo = FR_rsq * poly_lo + C_3
//      poly_hi = FR_rsq * poly_hi
//
        fma.s1	FR_Z = FR_Z, FR_rsq, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p9)  	fma.s1 FR_poly_hi = FR_rsq, FR_S_2, FR_S_1
      nop.i 999
}

{ .mfi
      nop.m 999
//
//      if (i_1 == 0):
//      poly_lo = FR_rsq * S_5 + S_4
//      poly_hi = FR_rsq * S_2 + S_1
//
(p10) 	fma.s1 FR_poly_lo = FR_rsq, FR_C_5, FR_C_4
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1 == 0):
//      Z = Z * r  for only one of the small r cases - not there
//      in original implementation notes.
//
(p9)  	fma.s1 FR_poly_lo = FR_rsq, FR_poly_lo, FR_S_3
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p10) 	fma.s1 FR_poly_hi = FR_rsq, FR_C_2, FR_C_1
      nop.i 999
}

{ .mfi
      nop.m 999
(p10)	fma.d.s0 FR_C_1 = FR_C_1, FR_C_1, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p9)  	fma.s1 FR_poly_hi = FR_poly_hi, FR_rsq, f0
      nop.i 999
}

{ .mfi
      nop.m 999
//
//      poly_lo = FR_rsq * poly_lo + S_3
//      poly_hi = FR_rsq * poly_hi
//
(p10) 	fma.s1 FR_poly_lo = FR_rsq, FR_poly_lo, FR_C_3
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p10) 	fma.s1 FR_poly_hi = FR_poly_hi, FR_rsq, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
// 	if (i_1 == 0): dummy fmpy's to flag inexact
// 	r = 1
//
(p9)	fma.s1 FR_poly_hi = FR_r, FR_poly_hi, f0
      nop.i 999
}

{ .mfi
      nop.m 999
//
// 	poly_hi = r * poly_hi
//
        fma.s1	FR_poly = FR_Z, FR_poly_lo, FR_c
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p12)	fms.s1	FR_r = f0, f1, FR_r
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      poly_hi = Z * poly_lo + c
// 	if i_0 == 1: r = -r
//
     	fma.s1	FR_poly = FR_poly, f1, FR_poly_hi
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p12)	fms.d.s0 FR_Input_X = FR_r, f1, FR_poly
      nop.i 999
}

{ .mfb
      nop.m 999
//
//      poly = poly + poly_hi
//
(p11)	fma.d.s0 FR_Input_X = FR_r, f1, FR_poly
//
//      if (i_0 == 0) Result = r + poly
//      if (i_0 != 0) Result = r - poly
//
       br.ret.sptk   b0 ;;
}
L(SINCOS_NORMAL_R):

{ .mii
      nop.m 999
    	extr.u	GR_i_1 = GR_N_Inc, 0, 1 ;;
//
//      Set table_ptr1 and table_ptr2 to base address of
//      constant table.
    	cmp.eq.unc p9, p10 = 0x0, GR_i_1 ;;
}

{ .mfi
      nop.m 999
    	fma.s1	FR_rsq = FR_r, FR_r, f0
    	extr.u	GR_i_0 = GR_N_Inc, 1, 1 ;;
}

{ .mfi
      nop.m 999
    	frcpa.s1 FR_r_hi, p6 = f1, FR_r
    	cmp.eq.unc p11, p12 = 0x0, GR_i_0
}
;;

// ******************************************************************
// ******************************************************************
// ******************************************************************
//
//      r and c have been computed.
//      We known whether this is the sine or cosine routine.
//      Make sure ftz mode is set - should be automatic when using wre
//      Get [i_0,i_1] - two lsb of N_fix_gr alone.
//

{ .mmi
      nop.m 999
      addl           GR_Table_Base   = @ltoff(FSINCOS_CONSTANTS#), gp
      nop.i 999
}
;;

{ .mmi
      ld8 GR_Table_Base = [GR_Table_Base]
      nop.m 999
      nop.i 999
}
;;


{ .mfi
(p10)	add GR_Table_Base = 384, GR_Table_Base
(p12)	fms.s1 FR_Input_X = f0, f1, f1
(p9)	add GR_Table_Base = 224, GR_Table_Base ;;
}

{ .mmf
      nop.m 999
(p10)	ldfe FR_QQ_8 = [GR_Table_Base], 16
//
//      if (i_1==0) poly = poly * FR_rsq + PP_1_lo
//      else        poly = FR_rsq * poly
//
(p11)	fma.s1 FR_Input_X = f0, f1, f1 ;;
}

{ .mmf
(p10)	ldfe FR_QQ_7 = [GR_Table_Base], 16
//
// 	Adjust table pointers based on i_0
//      Compute rsq = r * r
//
(p9)	ldfe FR_PP_8 = [GR_Table_Base], 16
    	fma.s1 FR_r_cubed = FR_r, FR_rsq, f0 ;;
}

{ .mmf
(p9)	ldfe FR_PP_7 = [GR_Table_Base], 16
(p10)	ldfe FR_QQ_6 = [GR_Table_Base], 16
//
//      Load PP_8 and QQ_8; PP_7 and QQ_7
//
    	frcpa.s1 FR_r_hi, p6 = f1, FR_r_hi ;;
}
//
//      if (i_1==0) poly =   PP_7 + FR_rsq * PP_8.
//      else        poly =   QQ_7 + FR_rsq * QQ_8.
//

{ .mmb
(p9)	ldfe FR_PP_6 = [GR_Table_Base], 16
(p10)	ldfe FR_QQ_5 = [GR_Table_Base], 16
      nop.b 999 ;;
}

{ .mmb
(p9)	ldfe FR_PP_5 = [GR_Table_Base], 16
(p10)	ldfe FR_S_1 = [GR_Table_Base], 16
      nop.b 999 ;;
}

{ .mmb
(p10)	ldfe FR_QQ_1 = [GR_Table_Base], 16
(p9)	ldfe FR_C_1 = [GR_Table_Base], 16
      nop.b 999 ;;
}

{ .mmi
(p10)	ldfe FR_QQ_4 = [GR_Table_Base], 16 ;;
(p9)	ldfe FR_PP_1 = [GR_Table_Base], 16
      nop.i 999 ;;
}

{ .mmf
(p10)	ldfe FR_QQ_3 = [GR_Table_Base], 16
//
//      if (i_1=0) corr = corr + c*c
//      else       corr = corr * c
//
(p9)	ldfe FR_PP_4 = [GR_Table_Base], 16
(p10)	fma.s1 FR_poly = FR_rsq, FR_QQ_8, FR_QQ_7 ;;
}
//
//      if (i_1=0) poly = rsq * poly + PP_5
//      else       poly = rsq * poly + QQ_5
//      Load PP_4 or QQ_4
//

{ .mmf
(p9)	ldfe FR_PP_3 = [GR_Table_Base], 16
(p10)	ldfe FR_QQ_2 = [GR_Table_Base], 16
//
//      r_hi =   frcpa(frcpa(r)).
//      r_cube = r * FR_rsq.
//
(p9)	fma.s1 FR_poly = FR_rsq, FR_PP_8, FR_PP_7 ;;
}
//
//      Do dummy multiplies so inexact is always set.
//

{ .mfi
(p9)	ldfe FR_PP_2 = [GR_Table_Base], 16
//
//      r_lo = r - r_hi
//
(p9)	fma.s1 FR_U_lo = FR_r_hi, FR_r_hi, f0
      nop.i 999 ;;
}

{ .mmf
      nop.m 999
(p9)	ldfe FR_PP_1_lo = [GR_Table_Base], 16
(p10)	fma.s1 FR_corr = FR_S_1, FR_r_cubed, FR_r
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_QQ_6
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1=0) U_lo = r_hi * r_hi
//      else       U_lo = r_hi + r
//
(p9)	fma.s1 FR_corr = FR_C_1, FR_rsq, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1=0) corr = C_1 * rsq
//      else       corr = S_1 * r_cubed + r
//
(p9)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_PP_6
      nop.i 999
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_U_lo = FR_r_hi, f1, FR_r
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1=0) U_hi = r_hi + U_hi
//      else       U_hi = QQ_1 * U_hi + 1
//
(p9)	fma.s1 FR_U_lo = FR_r, FR_r_hi, FR_U_lo
      nop.i 999
}

{ .mfi
      nop.m 999
//
//      U_hi = r_hi * r_hi
//
    	fms.s1 FR_r_lo = FR_r, f1, FR_r_hi
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      Load PP_1, PP_6, PP_5, and C_1
//      Load QQ_1, QQ_6, QQ_5, and S_1
//
    	fma.s1 FR_U_hi = FR_r_hi, FR_r_hi, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_QQ_5
      nop.i 999
}

{ .mfi
      nop.m 999
(p10)	fnma.s1	FR_corr = FR_corr, FR_c, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1=0) U_lo = r * r_hi + U_lo
//      else       U_lo = r_lo * U_lo
//
(p9)	fma.s1 FR_corr = FR_corr, FR_c, FR_c
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p9)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_PP_5
      nop.i 999
}

{ .mfi
      nop.m 999
//
//      if (i_1 =0) U_hi = r + U_hi
//      if (i_1 =0) U_lo = r_lo * U_lo
//
//
(p9)	fma.d.s0 FR_PP_5 = FR_PP_5, FR_PP_4, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p9)	fma.s1 FR_U_lo = FR_r, FR_r, FR_U_lo
      nop.i 999
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_U_lo = FR_r_lo, FR_U_lo, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1=0) poly = poly * rsq + PP_6
//      else       poly = poly * rsq + QQ_6
//
(p9)	fma.s1 FR_U_hi = FR_r_hi, FR_U_hi, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_QQ_4
      nop.i 999
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_U_hi = FR_QQ_1, FR_U_hi, f1
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p10)	fma.d.s0 FR_QQ_5 = FR_QQ_5, FR_QQ_5, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1!=0) U_hi = PP_1 * U_hi
//      if (i_1!=0) U_lo = r * r  + U_lo
//      Load PP_3 or QQ_3
//
(p9)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_PP_4
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p9)	fma.s1 FR_U_lo = FR_r_lo, FR_U_lo, f0
      nop.i 999
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_U_lo = FR_QQ_1,FR_U_lo, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p9)	fma.s1 FR_U_hi = FR_PP_1, FR_U_hi, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_QQ_3
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      Load PP_2, QQ_2
//
(p9)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_PP_3
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1==0) poly = FR_rsq * poly  + PP_3
//      else        poly = FR_rsq * poly  + QQ_3
//      Load PP_1_lo
//
(p9)	fma.s1 FR_U_lo = FR_PP_1, FR_U_lo, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1 =0) poly = poly * rsq + pp_r4
//      else        poly = poly * rsq + qq_r4
//
(p9)	fma.s1 FR_U_hi = FR_r, f1, FR_U_hi
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_QQ_2
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1==0) U_lo =  PP_1_hi * U_lo
//      else        U_lo =  QQ_1 * U_lo
//
(p9)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_PP_2
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_0==0)  Result = 1
//      else         Result = -1
//
     	fma.s1 FR_V = FR_U_lo, f1, FR_corr
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_poly = FR_rsq, FR_poly, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1==0) poly =  FR_rsq * poly + PP_2
//      else poly =  FR_rsq * poly + QQ_2
//
(p9)	fma.s1 FR_poly = FR_rsq, FR_poly, FR_PP_1_lo
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p10)	fma.s1 FR_poly = FR_rsq, FR_poly, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      V = U_lo + corr
//
(p9)	fma.s1 FR_poly = FR_r_cubed, FR_poly, f0
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
//
//      if (i_1==0) poly = r_cube * poly
//      else        poly = FR_rsq * poly
//
    	fma.s1	FR_V = FR_poly, f1, FR_V
      nop.i 999 ;;
}

{ .mfi
      nop.m 999
(p12)	fms.d.s0 FR_Input_X = FR_Input_X, FR_U_hi, FR_V
      nop.i 999
}

{ .mfb
      nop.m 999
//
//      V = V + poly
//
(p11)	fma.d.s0 FR_Input_X = FR_Input_X, FR_U_hi, FR_V
//
//      if (i_0==0) Result = Result * U_hi + V
//      else        Result = Result * U_hi - V
//
       br.ret.sptk   b0 ;;
}

//
//      If cosine, FR_Input_X = 1
//      If sine, FR_Input_X = +/-Zero (Input FR_Input_X)
//      Results are exact, no exceptions
//
L(SINCOS_ZERO):

{ .mmb
        cmp.eq.unc p6, p7 = 0x1, GR_Sin_or_Cos
      nop.m 999
      nop.b 999 ;;
}

{ .mfi
      nop.m 999
(p7)    fmerge.s FR_Input_X = FR_Input_X, FR_Input_X
      nop.i 999
}

{ .mfb
      nop.m 999
(p6)    fmerge.s FR_Input_X = f1, f1
       br.ret.sptk   b0 ;;
}

L(SINCOS_SPECIAL):

//
//      Path for Arg = +/- QNaN, SNaN, Inf
//      Invalid can be raised. SNaNs
//      become QNaNs
//

{ .mfb
      nop.m 999
        fmpy.d.s0 FR_Input_X = FR_Input_X, f0
        br.ret.sptk   b0 ;;
}
.endp __libm_cos_double_dbx#
ASM_SIZE_DIRECTIVE(__libm_cos_double_dbx#)



//
//      Call int pi_by_2_reduce(double* x, double *y)
//      for |arguments| >= 2**63
//      Address to save r and c as double
//
//
//      psp    sp+64
//             sp+48  -> f0 c
//      r45    sp+32  -> f0 r
//      r44 -> sp+16  -> InputX
//      sp     sp     -> scratch provided to callee



.proc __libm_callout_2
__libm_callout_2:
L(SINCOS_ARG_TOO_LARGE):

.prologue
{ .mfi
        add   r45=-32,sp                        // Parameter: r address
        nop.f 0
.save   ar.pfs,GR_SAVE_PFS
        mov  GR_SAVE_PFS=ar.pfs                 // Save ar.pfs
}
{ .mfi
.fframe 64
        add sp=-64,sp                           // Create new stack
        nop.f 0
        mov GR_SAVE_GP=gp                       // Save gp
};;
{ .mmi
        stfe [r45] = f0,16                      // Clear Parameter r on stack
        add  r44 = 16,sp                        // Parameter x address
.save   b0, GR_SAVE_B0
        mov GR_SAVE_B0=b0                       // Save b0
};;
.body
{ .mib
        stfe [r45] = f0,-16                     // Clear Parameter c on stack
        nop.i 0
        nop.b 0
}
{ .mib
        stfe [r44] = FR_Input_X                 // Store Parameter x on stack
        nop.i 0
        br.call.sptk b0=__libm_pi_by_2_reduce# ;;
};;


{ .mii
        ldfe  FR_Input_X =[r44],16
//
//      Get r and c off stack
//
        adds  GR_Table_Base1 = -16, GR_Table_Base1
//
//      Get r and c off stack
//
        add   GR_N_Inc = GR_Sin_or_Cos,r8 ;;
}
{ .mmb
        ldfe  FR_r =[r45],16
//
//      Get X off the stack
//      Readjust Table ptr
//
        ldfs FR_Two_to_M3 = [GR_Table_Base1],4
        nop.b 999 ;;
}
{ .mmb
        ldfs FR_Neg_Two_to_M3 = [GR_Table_Base1],0
        ldfe  FR_c =[r45]
        nop.b 999 ;;
}

{ .mfi
.restore sp
        add   sp = 64,sp                       // Restore stack pointer
        fcmp.lt.unc.s1  p6, p0 = FR_r, FR_Two_to_M3
        mov   b0 = GR_SAVE_B0                  // Restore return address
};;
{ .mib
        mov   gp = GR_SAVE_GP                  // Restore gp
        mov   ar.pfs = GR_SAVE_PFS             // Restore ar.pfs
        nop.b 0
};;


{ .mfi
      nop.m 999
(p6)    fcmp.gt.unc.s1	p6, p0 = FR_r, FR_Neg_Two_to_M3
      nop.i 999 ;;
}

{ .mib
      nop.m 999
      nop.i 999
(p6)    br.cond.spnt L(SINCOS_SMALL_R) ;;
}

{ .mib
      nop.m 999
      nop.i 999
        br.cond.sptk L(SINCOS_NORMAL_R) ;;
}

.endp __libm_callout_2
ASM_SIZE_DIRECTIVE(__libm_callout_2)

.type   __libm_pi_by_2_reduce#,@function
.global __libm_pi_by_2_reduce#


.type __libm_sin_double_dbx#,@function
.global __libm_sin_double_dbx#
.type __libm_cos_double_dbx#,@function
.global __libm_cos_double_dbx#