1 /* roundf4 - for each of four float slots, round to the nearest integer,
2 halfway cases are rounded away form zero.
3 Copyright (C) 2006, 2007 Sony Computer Entertainment Inc.
6 Redistribution and use in source and binary forms,
7 with or without modification, are permitted provided that the
8 following conditions are met:
9 * Redistributions of source code must retain the above copyright
10 notice, this list of conditions and the following disclaimer.
11 * Redistributions in binary form must reproduce the above copyright
12 notice, this list of conditions and the following disclaimer in the
13 documentation and/or other materials provided with the distribution.
14 * Neither the name of the Sony Computer Entertainment Inc nor the names
15 of its contributors may be used to endorse or promote products derived
16 from this software without specific prior written permission.
18 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
19 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21 ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
22 LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
23 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
24 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
25 INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
26 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
27 ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
28 POSSIBILITY OF SUCH DAMAGE.
31 #ifndef ___SIMD_MATH_ROUNDF4_H___
32 #define ___SIMD_MATH_ROUNDF4_H___
35 #include <spu_intrinsics.h>
37 static inline vector float
38 _roundf4(vector float in)
41 vec_uint4 or_mask, and_mask, mask, addend;
44 /* Add 0.5 (fixed precision to eliminate rounding issues
46 exp = spu_sub(125, spu_and(spu_rlmask((vec_int4)in, -23), 0xFF));
48 addend = spu_and(spu_rlmask( spu_splats((unsigned int)0x1000000), exp),
49 spu_cmpgt((vec_uint4)exp, -31));
51 in = (vec_float4)spu_add((vec_uint4)in, addend);
54 /* Truncate the result.
56 exp = spu_sub(127, spu_and(spu_rlmask((vec_int4)in, -23), 0xFF));
58 or_mask = spu_cmpgt(exp, 0);
60 and_mask = spu_rlmask(spu_splats((unsigned int)0x7FFFFF), exp);
62 mask = spu_or(spu_and(and_mask, spu_cmpgt(exp, -31)), or_mask);
64 out = spu_andc(in, (vec_float4)(mask));