1 /*
2 * Copyright © 2020 Microsoft Corporation
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_builder.h"
26
27 #include "util/u_math.h"
28
29 static bool
nir_scale_fdiv_instr(nir_builder * b,nir_instr * instr,UNUSED void * _data)30 nir_scale_fdiv_instr(nir_builder *b, nir_instr *instr, UNUSED void *_data)
31 {
32 if (instr->type != nir_instr_type_alu)
33 return false;
34
35 nir_alu_instr *alu = nir_instr_as_alu(instr);
36 if (alu->op != nir_op_fdiv || alu->src[0].src.ssa->bit_size != 32)
37 return false;
38
39 b->cursor = nir_before_instr(&alu->instr);
40
41 nir_def *orig_a = nir_ssa_for_alu_src(b, alu, 0);
42 nir_def *orig_b = nir_ssa_for_alu_src(b, alu, 1);
43 nir_def *fabs = nir_fabs(b, orig_b);
44 nir_def *big = nir_fgt_imm(b, fabs, uif(0x7e800000));
45 nir_def *small = nir_flt_imm(b, fabs, uif(0x00800000));
46
47 nir_def *scaled_down_a = nir_fmul_imm(b, orig_a, 0.25);
48 nir_def *scaled_down_b = nir_fmul_imm(b, orig_b, 0.25);
49 nir_def *scaled_up_a = nir_fmul_imm(b, orig_a, 16777216.0);
50 nir_def *scaled_up_b = nir_fmul_imm(b, orig_b, 16777216.0);
51
52 nir_def *final_a =
53 nir_bcsel(b, big, scaled_down_a,
54 (nir_bcsel(b, small, scaled_up_a, orig_a)));
55 nir_def *final_b =
56 nir_bcsel(b, big, scaled_down_b,
57 (nir_bcsel(b, small, scaled_up_b, orig_b)));
58
59 nir_def *new_div = nir_fdiv(b, final_a, final_b);
60 nir_def_rewrite_uses(&alu->def, new_div);
61
62 return true;
63 }
64
65 /** Scale both sides of an fdiv if needed to prevent denorm flushing
66 *
67 * This may be needed to satisfy the precision requirements of OpenCL. When
68 * fdiv is lowered to frcp+fmul, denorm flushing may cause the frcp to return
69 * zero even for finite floats. This multiplies both sides of an fdiv by a
70 * constant, if needed, to prevent such flushing.
71 */
72 bool
nir_scale_fdiv(nir_shader * shader)73 nir_scale_fdiv(nir_shader *shader)
74 {
75 return nir_shader_instructions_pass(shader, nir_scale_fdiv_instr,
76 nir_metadata_block_index |
77 nir_metadata_dominance,
78 NULL);
79 }
80