1 /*
2 (C) Copyright 2001,2006,
3 International Business Machines Corporation,
4 Sony Computer Entertainment, Incorporated,
5 Toshiba Corporation,
6
7 All rights reserved.
8
9 Redistribution and use in source and binary forms, with or without
10 modification, are permitted provided that the following conditions are met:
11
12 * Redistributions of source code must retain the above copyright notice,
13 this list of conditions and the following disclaimer.
14 * Redistributions in binary form must reproduce the above copyright
15 notice, this list of conditions and the following disclaimer in the
16 documentation and/or other materials provided with the distribution.
17 * Neither the names of the copyright holders nor the names of their
18 contributors may be used to endorse or promote products derived from this
19 software without specific prior written permission.
20
21 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS
22 IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
23 TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
24 PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
25 OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
26 EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
27 PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
28 PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
29 LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
30 NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
31 SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
32 */
33 #ifndef _FLOOR_H_
34 #define _FLOOR_H_ 1
35
36 #include <spu_intrinsics.h>
37 #include "headers/vec_literal.h"
38
39 /* Round the input downwards to the nearest integer.
40 */
_floor(double x)41 static __inline double _floor(double x)
42 {
43 vec_uchar16 swap_words = VEC_LITERAL(vec_uchar16, 4,5,6,7, 0,1,2,3, 12,13,14,15, 8,9,10,11);
44 vec_uchar16 splat_hi = VEC_LITERAL(vec_uchar16, 0,1,2,3,0,1,2,3, 8,9,10,11, 8,9,10,11);
45 vec_uint4 one = VEC_LITERAL(vec_uint4, 0, 1, 0, 1);
46 vec_int4 exp, shift;
47 vec_uint4 mask, mask_1, frac_mask, addend, insert, pos, equal0;
48 vec_ullong2 sign = VEC_SPLAT_U64(0x8000000000000000ULL);
49 vec_double2 in, in_hi, out;
50 vec_double2 minus_one = VEC_SPLAT_F64(-1.0);
51
52 in = spu_promote(x, 0);
53
54 /* This function generates the following component
55 * based upon the inputs.
56 *
57 * mask = bits of the input that need to be replaced.
58 * insert = value of the bits that need to be replaced
59 * addend = value to be added to perform function.
60 *
61 * These are applied as follows:.
62 *
63 * out = ((in & mask) | insert) + addend
64 */
65 in_hi = spu_shuffle(in, in, splat_hi);
66 pos = spu_cmpgt((vec_int4)in_hi, -1);
67 exp = spu_and(spu_rlmask((vec_int4)in_hi, -20), 0x7FF);
68 shift = spu_sub(VEC_LITERAL(vec_int4, 1023, 1043, 1023, 1043), exp);
69
70 /* clamp shift to the range 0 to -31.
71 */
72 shift = spu_sel(VEC_SPLAT_S32(-32), spu_andc(shift, (vec_int4)spu_cmpgt(shift, 0)), spu_cmpgt(shift, -32));
73
74 frac_mask = spu_rlmask(VEC_LITERAL(vec_uint4, 0xFFFFF, -1, 0xFFFFF, -1), shift);
75 mask = spu_orc(frac_mask, spu_cmpgt(exp, 0x3FE));
76
77 /* addend = ((in & mask) && (in >= 0)) ? mask+1 : 0
78 */
79 mask_1 = spu_addx(mask, one, spu_rlqwbyte(spu_genc(mask, one), 4));
80
81 equal0 = spu_cmpeq(spu_and((vec_uint4)in, mask), 0);
82 addend = spu_andc(spu_andc(mask_1, pos),
83 spu_and(equal0, spu_shuffle(equal0, equal0, swap_words)));
84
85 insert = spu_andc(spu_andc((vec_uint4)minus_one, pos),
86 spu_cmpgt((vec_uint4)spu_add(exp, -1), 1022));
87
88 in = spu_sel(in, (vec_double2)insert, spu_andc((vec_ullong2)mask, sign));
89 out = (vec_double2)spu_addx((vec_uint4)in, addend,
90 spu_rlqwbyte(spu_genc((vec_uint4)in, addend), 4));
91
92 return (spu_extract(out, 0));
93 }
94 #endif /* _FLOOR_H_ */
95
