1/* 2 * Copyright © 2012, 2013, 2014 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 DEALINGS 21 * IN THE SOFTWARE. 22 */ 23 24#include "brw_vec4.h" 25#include "brw_vec4_live_variables.h" 26#include "brw_cfg.h" 27 28using namespace brw; 29 30/** @file brw_vec4_cse.cpp 31 * 32 * Support for local common subexpression elimination. 33 * 34 * See Muchnick's Advanced Compiler Design and Implementation, section 35 * 13.1 (p378). 36 */ 37 38namespace { 39struct aeb_entry : public exec_node { 40 /** The instruction that generates the expression value. */ 41 vec4_instruction *generator; 42 43 /** The temporary where the value is stored. */ 44 src_reg tmp; 45}; 46} 47 48static bool 49is_expression(const vec4_instruction *const inst) 50{ 51 switch (inst->opcode) { 52 case BRW_OPCODE_MOV: 53 case BRW_OPCODE_SEL: 54 case BRW_OPCODE_NOT: 55 case BRW_OPCODE_AND: 56 case BRW_OPCODE_OR: 57 case BRW_OPCODE_XOR: 58 case BRW_OPCODE_SHR: 59 case BRW_OPCODE_SHL: 60 case BRW_OPCODE_ASR: 61 case BRW_OPCODE_CMP: 62 case BRW_OPCODE_CMPN: 63 case BRW_OPCODE_ADD: 64 case BRW_OPCODE_MUL: 65 case SHADER_OPCODE_MULH: 66 case BRW_OPCODE_FRC: 67 case BRW_OPCODE_RNDU: 68 case BRW_OPCODE_RNDD: 69 case BRW_OPCODE_RNDE: 70 case BRW_OPCODE_RNDZ: 71 case BRW_OPCODE_LINE: 72 case BRW_OPCODE_PLN: 73 case BRW_OPCODE_MAD: 74 case BRW_OPCODE_LRP: 75 case VEC4_OPCODE_UNPACK_UNIFORM: 76 case SHADER_OPCODE_FIND_LIVE_CHANNEL: 77 case SHADER_OPCODE_BROADCAST: 78 case VEC4_TCS_OPCODE_SET_INPUT_URB_OFFSETS: 79 case VEC4_TCS_OPCODE_SET_OUTPUT_URB_OFFSETS: 80 return true; 81 case SHADER_OPCODE_RCP: 82 case SHADER_OPCODE_RSQ: 83 case SHADER_OPCODE_SQRT: 84 case SHADER_OPCODE_EXP2: 85 case SHADER_OPCODE_LOG2: 86 case SHADER_OPCODE_POW: 87 case SHADER_OPCODE_INT_QUOTIENT: 88 case SHADER_OPCODE_INT_REMAINDER: 89 case SHADER_OPCODE_SIN: 90 case SHADER_OPCODE_COS: 91 return inst->mlen == 0; 92 default: 93 return false; 94 } 95} 96 97static bool 98operands_match(const vec4_instruction *a, const vec4_instruction *b) 99{ 100 const src_reg *xs = a->src; 101 const src_reg *ys = b->src; 102 103 if (a->opcode == BRW_OPCODE_MAD) { 104 return xs[0].equals(ys[0]) && 105 ((xs[1].equals(ys[1]) && xs[2].equals(ys[2])) || 106 (xs[2].equals(ys[1]) && xs[1].equals(ys[2]))); 107 } else if (a->opcode == BRW_OPCODE_MOV && 108 xs[0].file == IMM && 109 xs[0].type == BRW_REGISTER_TYPE_VF) { 110 src_reg tmp_x = xs[0]; 111 src_reg tmp_y = ys[0]; 112 113 /* Smash out the values that are not part of the writemask. Otherwise 114 * the equals operator will fail due to mismatches in unused components. 115 */ 116 const unsigned ab_writemask = a->dst.writemask & b->dst.writemask; 117 const uint32_t mask = ((ab_writemask & WRITEMASK_X) ? 0x000000ff : 0) | 118 ((ab_writemask & WRITEMASK_Y) ? 0x0000ff00 : 0) | 119 ((ab_writemask & WRITEMASK_Z) ? 0x00ff0000 : 0) | 120 ((ab_writemask & WRITEMASK_W) ? 0xff000000 : 0); 121 122 tmp_x.ud &= mask; 123 tmp_y.ud &= mask; 124 125 return tmp_x.equals(tmp_y); 126 } else if (!a->is_commutative()) { 127 return xs[0].equals(ys[0]) && xs[1].equals(ys[1]) && xs[2].equals(ys[2]); 128 } else { 129 return (xs[0].equals(ys[0]) && xs[1].equals(ys[1])) || 130 (xs[1].equals(ys[0]) && xs[0].equals(ys[1])); 131 } 132} 133 134/** 135 * Checks if instructions match, exactly for sources, but loosely for 136 * destination writemasks. 137 * 138 * \param 'a' is the generating expression from the AEB entry. 139 * \param 'b' is the second occurrence of the expression that we're 140 * considering eliminating. 141 */ 142static bool 143instructions_match(vec4_instruction *a, vec4_instruction *b) 144{ 145 return a->opcode == b->opcode && 146 a->saturate == b->saturate && 147 a->predicate == b->predicate && 148 a->predicate_inverse == b->predicate_inverse && 149 a->conditional_mod == b->conditional_mod && 150 a->flag_subreg == b->flag_subreg && 151 a->dst.type == b->dst.type && 152 a->offset == b->offset && 153 a->mlen == b->mlen && 154 a->base_mrf == b->base_mrf && 155 a->header_size == b->header_size && 156 a->shadow_compare == b->shadow_compare && 157 ((a->dst.writemask & b->dst.writemask) == a->dst.writemask) && 158 a->force_writemask_all == b->force_writemask_all && 159 a->size_written == b->size_written && 160 a->exec_size == b->exec_size && 161 a->group == b->group && 162 operands_match(a, b); 163} 164 165bool 166vec4_visitor::opt_cse_local(bblock_t *block, const vec4_live_variables &live) 167{ 168 bool progress = false; 169 exec_list aeb; 170 171 void *cse_ctx = ralloc_context(NULL); 172 173 int ip = block->start_ip; 174 foreach_inst_in_block (vec4_instruction, inst, block) { 175 /* Skip some cases. */ 176 if (is_expression(inst) && !inst->predicate && inst->mlen == 0 && 177 ((inst->dst.file != ARF && inst->dst.file != FIXED_GRF) || 178 inst->dst.is_null())) 179 { 180 bool found = false; 181 182 foreach_in_list_use_after(aeb_entry, entry, &aeb) { 183 /* Match current instruction's expression against those in AEB. */ 184 if (!(entry->generator->dst.is_null() && !inst->dst.is_null()) && 185 instructions_match(inst, entry->generator)) { 186 found = true; 187 progress = true; 188 break; 189 } 190 } 191 192 if (!found) { 193 if (inst->opcode != BRW_OPCODE_MOV || 194 (inst->opcode == BRW_OPCODE_MOV && 195 inst->src[0].file == IMM && 196 inst->src[0].type == BRW_REGISTER_TYPE_VF)) { 197 /* Our first sighting of this expression. Create an entry. */ 198 aeb_entry *entry = ralloc(cse_ctx, aeb_entry); 199 entry->tmp = src_reg(); /* file will be BAD_FILE */ 200 entry->generator = inst; 201 aeb.push_tail(entry); 202 } 203 } else { 204 /* This is at least our second sighting of this expression. 205 * If we don't have a temporary already, make one. 206 */ 207 bool no_existing_temp = entry->tmp.file == BAD_FILE; 208 if (no_existing_temp && !entry->generator->dst.is_null()) { 209 entry->tmp = retype(src_reg(VGRF, alloc.allocate( 210 regs_written(entry->generator)), 211 NULL), inst->dst.type); 212 213 const unsigned width = entry->generator->exec_size; 214 unsigned component_size = width * type_sz(entry->tmp.type); 215 unsigned num_copy_movs = 216 DIV_ROUND_UP(entry->generator->size_written, component_size); 217 for (unsigned i = 0; i < num_copy_movs; ++i) { 218 vec4_instruction *copy = 219 MOV(offset(entry->generator->dst, width, i), 220 offset(entry->tmp, width, i)); 221 copy->exec_size = width; 222 copy->group = entry->generator->group; 223 copy->force_writemask_all = 224 entry->generator->force_writemask_all; 225 entry->generator->insert_after(block, copy); 226 } 227 228 entry->generator->dst = dst_reg(entry->tmp); 229 } 230 231 /* dest <- temp */ 232 if (!inst->dst.is_null()) { 233 assert(inst->dst.type == entry->tmp.type); 234 const unsigned width = inst->exec_size; 235 unsigned component_size = width * type_sz(inst->dst.type); 236 unsigned num_copy_movs = 237 DIV_ROUND_UP(inst->size_written, component_size); 238 for (unsigned i = 0; i < num_copy_movs; ++i) { 239 vec4_instruction *copy = 240 MOV(offset(inst->dst, width, i), 241 offset(entry->tmp, width, i)); 242 copy->exec_size = inst->exec_size; 243 copy->group = inst->group; 244 copy->force_writemask_all = inst->force_writemask_all; 245 inst->insert_before(block, copy); 246 } 247 } 248 249 /* Set our iterator so that next time through the loop inst->next 250 * will get the instruction in the basic block after the one we've 251 * removed. 252 */ 253 vec4_instruction *prev = (vec4_instruction *)inst->prev; 254 255 inst->remove(block); 256 inst = prev; 257 } 258 } 259 260 foreach_in_list_safe(aeb_entry, entry, &aeb) { 261 /* Kill all AEB entries that write a different value to or read from 262 * the flag register if we just wrote it. 263 */ 264 if (inst->writes_flag(devinfo)) { 265 if (entry->generator->reads_flag() || 266 (entry->generator->writes_flag(devinfo) && 267 !instructions_match(inst, entry->generator))) { 268 entry->remove(); 269 ralloc_free(entry); 270 continue; 271 } 272 } 273 274 for (int i = 0; i < 3; i++) { 275 src_reg *src = &entry->generator->src[i]; 276 277 /* Kill all AEB entries that use the destination we just 278 * overwrote. 279 */ 280 if (inst->dst.file == entry->generator->src[i].file && 281 inst->dst.nr == entry->generator->src[i].nr) { 282 entry->remove(); 283 ralloc_free(entry); 284 break; 285 } 286 287 /* Kill any AEB entries using registers that don't get reused any 288 * more -- a sure sign they'll fail operands_match(). 289 */ 290 if (src->file == VGRF) { 291 if (live.var_range_end(var_from_reg(alloc, dst_reg(*src)), 8) < ip) { 292 entry->remove(); 293 ralloc_free(entry); 294 break; 295 } 296 } 297 } 298 } 299 300 ip++; 301 } 302 303 ralloc_free(cse_ctx); 304 305 return progress; 306} 307 308bool 309vec4_visitor::opt_cse() 310{ 311 bool progress = false; 312 const vec4_live_variables &live = live_analysis.require(); 313 314 foreach_block (block, cfg) { 315 progress = opt_cse_local(block, live) || progress; 316 } 317 318 if (progress) 319 invalidate_analysis(DEPENDENCY_INSTRUCTIONS | DEPENDENCY_VARIABLES); 320 321 return progress; 322} 323