src/amd/common/ac_lower_subgroups.c

   1 /*
   2  * Copyright © 2018 Google 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/nir.h"
  25 #include "nir/nir_builder.h"
  26
  27 #include "ac_nir_to_llvm.h"
  28
  29 static nir_ssa_def *ac_lower_subgroups_intrin(nir_builder *b, nir_intrinsic_instr *intrin)
  30 {
  31         switch(intrin->intrinsic) {
  32         case nir_intrinsic_vote_ieq:
  33         case nir_intrinsic_vote_feq: {
  34                 nir_intrinsic_instr *rfi =
  35                         nir_intrinsic_instr_create(b->shader, nir_intrinsic_read_first_invocation);
  36                 nir_ssa_dest_init(&rfi->instr, &rfi->dest,
  37                                   1, intrin->src[0].ssa->bit_size, NULL);
  38                 nir_src_copy(&rfi->src[0], &intrin->src[0], rfi);
  39                 rfi->num_components = 1;
  40
  41                 nir_ssa_def *is_ne;
  42                 if (intrin->intrinsic == nir_intrinsic_vote_feq)
  43                         is_ne = nir_fne(b, &rfi->dest.ssa, intrin->src[0].ssa);
  44                 else
  45                         is_ne = nir_ine(b, &rfi->dest.ssa, intrin->src[0].ssa);
  46
  47                 nir_intrinsic_instr *ballot =
  48                         nir_intrinsic_instr_create(b->shader, nir_intrinsic_ballot);
  49                 nir_ssa_dest_init(&ballot->instr, &ballot->dest,
  50                                   1, 64, NULL);
  51                 ballot->src[0] = nir_src_for_ssa(is_ne);
  52                 ballot->num_components = 1;
  53
  54                 return nir_ieq(b, &ballot->dest.ssa, nir_imm_int64(b, 0));
  55         }
  56         default:
  57                 return NULL;
  58         }
  59 }
  60
  61 bool ac_lower_subgroups(struct nir_shader *shader)
  62 {
  63         bool progress = false;
  64
  65         nir_foreach_function(function, shader) {
  66                 if (!function->impl)
  67                         continue;
  68
  69                 nir_builder b;
  70                 nir_builder_init(&b, function->impl);
  71
  72                 nir_foreach_block(block, function->impl) {
  73                         nir_foreach_instr_safe(instr, block) {
  74                                 if (instr->type != nir_instr_type_intrinsic)
  75                                         continue;
  76
  77                                 nir_intrinsic_instr *intrin = nir_instr_as_intrinsic(instr);
  78                                 b.cursor = nir_before_instr(instr);
  79
  80                                 nir_ssa_def *lower = ac_lower_subgroups_intrin(&b, intrin);
  81                                 if (!lower)
  82                                         continue;
  83
  84                                 nir_ssa_def_rewrite_uses(&intrin->dest.ssa, nir_src_for_ssa(lower));
  85                                 nir_instr_remove(instr);
  86                                 progress = true;
  87                         }
  88                 }
  89         }
  90
  91         return progress;
  92 }