1 /* -------------------------------------------------------------- */
2 /* (C)Copyright 2001,2008, */
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 */
10 /* without modification, are permitted provided that the */
11 /* following conditions are met: */
12 /* */
13 /* - Redistributions of source code must retain the above copyright*/
14 /* notice, this list of conditions and the following disclaimer. */
15 /* */
16 /* - Redistributions in binary form must reproduce the above */
17 /* copyright notice, this list of conditions and the following */
18 /* disclaimer in the documentation and/or other materials */
19 /* provided with the distribution. */
20 /* */
21 /* - Neither the name of IBM Corporation nor the names of its */
22 /* contributors may be used to endorse or promote products */
23 /* derived from this software without specific prior written */
24 /* permission. */
25 /* */
26 /* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND */
27 /* CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, */
28 /* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF */
29 /* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE */
30 /* DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR */
31 /* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, */
32 /* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT */
33 /* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; */
34 /* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) */
35 /* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN */
36 /* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR */
37 /* OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, */
38 /* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
39 /* -------------------------------------------------------------- */
40 /* PROLOG END TAG zYx */
41 #ifdef __SPU__
42
43 #ifndef _FLOORD2_H_
44 #define _FLOORD2_H_ 1
45
46 #include <spu_intrinsics.h>
47
48
49 /*
50 * FUNCTION
51 * vector double _floord2(vector double x)
52 *
53 * DESCRIPTION
54 * The _floord2 function rounds the elements of an vector double
55 * input vector downwards to their nearest integer representable
56 * as a double.
57 *
58 */
_floord2(vector double in)59 static __inline vector double _floord2(vector double in)
60 {
61 vec_uchar16 swap_words = (vec_uchar16) { 4,5,6,7, 0,1,2,3, 12,13,14,15, 8,9,10,11 };
62 vec_uchar16 splat_hi = (vec_uchar16) { 0,1,2,3, 0,1,2,3, 8,9,10,11, 8,9,10,11 };
63 vec_uint4 one = (vec_uint4) { 0, 1, 0, 1 };
64 vec_int4 exp, shift;
65 vec_uint4 mask, mask_1, frac_mask, addend, insert, pos, equal0;
66 vec_ullong2 sign = spu_splats(0x8000000000000000ULL);
67 vec_double2 in_hi, out;
68 vec_double2 minus_one = spu_splats(-1.0);
69
70 /* This function generates the following component
71 * based upon the inputs.
72 *
73 * mask = bits of the input that need to be replaced.
74 * insert = value of the bits that need to be replaced
75 * addend = value to be added to perform function.
76 *
77 * These are applied as follows:.
78 *
79 * out = ((in & mask) | insert) + addend
80 */
81 in_hi = spu_shuffle(in, in, splat_hi);
82 pos = spu_cmpgt((vec_int4)in_hi, -1);
83 exp = spu_and(spu_rlmask((vec_int4)in_hi, -20), 0x7FF);
84 shift = spu_sub(((vec_int4) { 1023, 1043, 1023, 1043 } ), exp);
85
86 /* clamp shift to the range 0 to -31.
87 */
88 shift = spu_sel(spu_splats(-32), spu_andc(shift, (vec_int4)spu_cmpgt(shift, 0)), spu_cmpgt(shift, -32));
89
90 frac_mask = spu_rlmask(((vec_uint4) { 0xFFFFF, -1, 0xFFFFF, -1 } ), shift);
91 mask = spu_orc(frac_mask, spu_cmpgt(exp, 0x3FE));
92
93 /* addend = ((in & mask) && (in >= 0)) ? mask+1 : 0
94 */
95 mask_1 = spu_addx(mask, one, spu_rlqwbyte(spu_genc(mask, one), 4));
96
97 equal0 = spu_cmpeq(spu_and((vec_uint4)in, mask), 0);
98 addend = spu_andc(spu_andc(mask_1, pos), spu_and(equal0, spu_shuffle(equal0, equal0, swap_words)));
99
100 insert = spu_andc(spu_andc((vec_uint4)minus_one, pos),
101 spu_cmpgt((vec_uint4)spu_add(exp, -1), 1022));
102
103 in = spu_sel(in, (vec_double2)insert, spu_andc((vec_ullong2)mask, sign));
104 out = (vec_double2)spu_addx((vec_uint4)in, addend, spu_rlqwbyte(spu_genc((vec_uint4)in, addend), 4));
105
106 return (out);
107 }
108
109 #endif /* _FLOORD2_H_ */
110 #endif /* __SPU__ */
111
