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 _LLROUND_H_
34 #define _LLROUND_H_ 1
35
36 #include <spu_intrinsics.h>
37 #include "headers/vec_literal.h"
38
_llround(double x)39 static __inline long long int _llround(double x)
40 {
41 int shift;
42 vec_int4 exp;
43 vec_uint4 mant, sign, mask, borrow, addend;
44 vec_double2 in;
45
46 in = spu_promote(x, 0);
47
48 /* Determine how many bits to shift the mantissa to correctly
49 * align it into long long element 0.
50 */
51 exp = spu_and(spu_rlmask((vec_int4)in, -20), 0x7FF);
52 exp = spu_add(exp, -1011);
53 shift = spu_extract(exp, 0);
54
55 mask = spu_cmpgt(exp, 0);
56 mask = (vec_uint4)spu_maskw(spu_extract(mask, 0));
57
58 /* Algn mantissa bits
59 */
60 mant = spu_sel(spu_rlmaskqwbyte((vec_uint4)in, -8), VEC_SPLAT_U32(0x00100000),
61 VEC_LITERAL(vec_uint4, 0,0,0xFFF00000,0));
62
63 mant = spu_and(spu_slqwbytebc(spu_slqw(mant, shift), shift), mask);
64
65 /* Perform round by adding 1 if the fraction bits are
66 * greater than or equal to .5
67 */
68 addend = spu_and(spu_rlqw(mant, 1), VEC_LITERAL(vec_uint4, 0,1,0,0));
69 mant = spu_addx(mant, addend, spu_rlqwbyte(spu_genc(mant, addend), 4));
70
71 /* Compute the two's complement of the mantissa if the
72 * input is negative.
73 */
74 sign = spu_maskw(spu_extract(spu_rlmaska((vec_int4)in, -31), 0));
75
76 mant = spu_xor(mant, sign);
77 borrow = spu_genb(mant, sign);
78 borrow = spu_shuffle(borrow, borrow,
79 VEC_LITERAL(vec_uchar16, 4,5,6,7, 192,192,192,192,
80 4,5,6,7, 192,192,192,192));
81 mant = spu_subx(mant, sign, borrow);
82
83 return (spu_extract((vec_llong2)(mant), 0));
84 }
85 #endif /* _LLROUND_H_ */
86