nir/algebraic: Simplify fsat of fsign
[mesa.git] / src / compiler / nir / nir_builtin_builder.c
1 /*
2 * Copyright © 2018 Red Hat Inc.
3 *
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8 * and/or sell copies of the Software, and to permit persons to whom the
9 * Software is furnished to do so, subject to the following conditions:
10 *
11 * The above copyright notice and this permission notice (including the next
12 * paragraph) shall be included in all copies or substantial portions of the
13 * Software.
14 *
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
18 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
20 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
21 * IN THE SOFTWARE.
22 */
23
24 #include "nir.h"
25 #include "nir_builtin_builder.h"
26
27 nir_ssa_def*
28 nir_cross(nir_builder *b, nir_ssa_def *x, nir_ssa_def *y)
29 {
30 unsigned yzx[3] = { 1, 2, 0 };
31 unsigned zxy[3] = { 2, 0, 1 };
32
33 return nir_fsub(b, nir_fmul(b, nir_swizzle(b, x, yzx, 3, true),
34 nir_swizzle(b, y, zxy, 3, true)),
35 nir_fmul(b, nir_swizzle(b, x, zxy, 3, true),
36 nir_swizzle(b, y, yzx, 3, true)));
37 }
38
39 nir_ssa_def*
40 nir_fast_length(nir_builder *b, nir_ssa_def *vec)
41 {
42 switch (vec->num_components) {
43 case 1: return nir_fsqrt(b, nir_fmul(b, vec, vec));
44 case 2: return nir_fsqrt(b, nir_fdot2(b, vec, vec));
45 case 3: return nir_fsqrt(b, nir_fdot3(b, vec, vec));
46 case 4: return nir_fsqrt(b, nir_fdot4(b, vec, vec));
47 default:
48 unreachable("Invalid number of components");
49 }
50 }
51
52 nir_ssa_def*
53 nir_smoothstep(nir_builder *b, nir_ssa_def *edge0, nir_ssa_def *edge1, nir_ssa_def *x)
54 {
55 nir_ssa_def *f2 = nir_imm_floatN_t(b, 2.0, x->bit_size);
56 nir_ssa_def *f3 = nir_imm_floatN_t(b, 3.0, x->bit_size);
57
58 /* t = clamp((x - edge0) / (edge1 - edge0), 0, 1) */
59 nir_ssa_def *t =
60 nir_fsat(b, nir_fdiv(b, nir_fsub(b, x, edge0),
61 nir_fsub(b, edge1, edge0)));
62
63 /* result = t * t * (3 - 2 * t) */
64 return nir_fmul(b, t, nir_fmul(b, t, nir_fsub(b, f3, nir_fmul(b, f2, t))));
65 }