0c3a6383190bcc7e020512d023e8a1465db1f8a9
[mesa.git] / src / glsl / ir_optimization.h
1 /*
2 * Copyright © 2010 Intel 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
21 * DEALINGS IN THE SOFTWARE.
22 */
23
24
25 /**
26 * \file ir_optimization.h
27 *
28 * Prototypes for optimization passes to be called by the compiler and drivers.
29 */
30
31 /* Operations for lower_instructions() */
32 #define SUB_TO_ADD_NEG 0x01
33 #define DIV_TO_MUL_RCP 0x02
34 #define EXP_TO_EXP2 0x04
35 #define POW_TO_EXP2 0x08
36 #define LOG_TO_LOG2 0x10
37 #define MOD_TO_FRACT 0x20
38 #define INT_DIV_TO_MUL_RCP 0x40
39 #define BITFIELD_INSERT_TO_BFM_BFI 0x80
40 #define LDEXP_TO_ARITH 0x100
41 #define CARRY_TO_ARITH 0x200
42 #define BORROW_TO_ARITH 0x400
43 #define SAT_TO_CLAMP 0x800
44
45 /**
46 * \see class lower_packing_builtins_visitor
47 */
48 enum lower_packing_builtins_op {
49 LOWER_PACK_UNPACK_NONE = 0x0000,
50
51 LOWER_PACK_SNORM_2x16 = 0x0001,
52 LOWER_UNPACK_SNORM_2x16 = 0x0002,
53
54 LOWER_PACK_UNORM_2x16 = 0x0004,
55 LOWER_UNPACK_UNORM_2x16 = 0x0008,
56
57 LOWER_PACK_HALF_2x16 = 0x0010,
58 LOWER_UNPACK_HALF_2x16 = 0x0020,
59
60 LOWER_PACK_HALF_2x16_TO_SPLIT = 0x0040,
61 LOWER_UNPACK_HALF_2x16_TO_SPLIT = 0x0080,
62
63 LOWER_PACK_SNORM_4x8 = 0x0100,
64 LOWER_UNPACK_SNORM_4x8 = 0x0200,
65
66 LOWER_PACK_UNORM_4x8 = 0x0400,
67 LOWER_UNPACK_UNORM_4x8 = 0x0800
68 };
69
70 bool do_common_optimization(exec_list *ir, bool linked,
71 bool uniform_locations_assigned,
72 const struct gl_shader_compiler_options *options,
73 bool native_integers);
74
75 bool do_rebalance_tree(exec_list *instructions);
76 bool do_algebraic(exec_list *instructions, bool native_integers,
77 const struct gl_shader_compiler_options *options);
78 bool do_constant_folding(exec_list *instructions);
79 bool do_constant_variable(exec_list *instructions);
80 bool do_constant_variable_unlinked(exec_list *instructions);
81 bool do_copy_propagation(exec_list *instructions);
82 bool do_copy_propagation_elements(exec_list *instructions);
83 bool do_constant_propagation(exec_list *instructions);
84 bool do_cse(exec_list *instructions);
85 void do_dead_builtin_varyings(struct gl_context *ctx,
86 gl_shader *producer, gl_shader *consumer,
87 unsigned num_tfeedback_decls,
88 class tfeedback_decl *tfeedback_decls);
89 bool do_dead_code(exec_list *instructions, bool uniform_locations_assigned);
90 bool do_dead_code_local(exec_list *instructions);
91 bool do_dead_code_unlinked(exec_list *instructions);
92 bool do_dead_functions(exec_list *instructions);
93 bool opt_flip_matrices(exec_list *instructions);
94 bool do_function_inlining(exec_list *instructions);
95 bool do_lower_jumps(exec_list *instructions, bool pull_out_jumps = true, bool lower_sub_return = true, bool lower_main_return = false, bool lower_continue = false, bool lower_break = false);
96 bool do_lower_texture_projection(exec_list *instructions);
97 bool do_if_simplification(exec_list *instructions);
98 bool opt_flatten_nested_if_blocks(exec_list *instructions);
99 bool do_discard_simplification(exec_list *instructions);
100 bool lower_if_to_cond_assign(exec_list *instructions, unsigned max_depth = 0);
101 bool do_mat_op_to_vec(exec_list *instructions);
102 bool do_noop_swizzle(exec_list *instructions);
103 bool do_structure_splitting(exec_list *instructions);
104 bool do_swizzle_swizzle(exec_list *instructions);
105 bool do_vectorize(exec_list *instructions);
106 bool do_tree_grafting(exec_list *instructions);
107 bool do_vec_index_to_cond_assign(exec_list *instructions);
108 bool do_vec_index_to_swizzle(exec_list *instructions);
109 bool lower_discard(exec_list *instructions);
110 void lower_discard_flow(exec_list *instructions);
111 bool lower_instructions(exec_list *instructions, unsigned what_to_lower);
112 bool lower_noise(exec_list *instructions);
113 bool lower_variable_index_to_cond_assign(exec_list *instructions,
114 bool lower_input, bool lower_output, bool lower_temp, bool lower_uniform);
115 bool lower_quadop_vector(exec_list *instructions, bool dont_lower_swz);
116 bool lower_clip_distance(gl_shader *shader);
117 void lower_output_reads(exec_list *instructions);
118 bool lower_packing_builtins(exec_list *instructions, int op_mask);
119 void lower_ubo_reference(struct gl_shader *shader, exec_list *instructions);
120 void lower_packed_varyings(void *mem_ctx,
121 unsigned locations_used, ir_variable_mode mode,
122 unsigned gs_input_vertices, gl_shader *shader);
123 bool lower_vector_insert(exec_list *instructions, bool lower_nonconstant_index);
124 void lower_named_interface_blocks(void *mem_ctx, gl_shader *shader);
125 bool optimize_redundant_jumps(exec_list *instructions);
126 bool optimize_split_arrays(exec_list *instructions, bool linked);
127 bool lower_offset_arrays(exec_list *instructions);
128 void optimize_dead_builtin_variables(exec_list *instructions,
129 enum ir_variable_mode other);
130
131 bool lower_vertex_id(gl_shader *shader);
132
133 ir_rvalue *
134 compare_index_block(exec_list *instructions, ir_variable *index,
135 unsigned base, unsigned components, void *mem_ctx);