diff options
author | Jason Ekstrand <[email protected]> | 2014-11-04 10:12:14 -0800 |
---|---|---|
committer | Jason Ekstrand <[email protected]> | 2015-01-15 07:19:01 -0800 |
commit | 13ec15bdbf4fa7e9681b46a4636575f1e4a700b4 (patch) | |
tree | f2e0d98cae2e701099a2cc1773b9c0d43516c8f4 | |
parent | ef7ebb908ed0ebd005bb5eb4554dc2ba8d6be983 (diff) |
nir: Add a peephole select optimization
Reviewed-by: Connor Abbott <[email protected]>
-rw-r--r-- | src/glsl/Makefile.sources | 1 | ||||
-rw-r--r-- | src/glsl/nir/nir.h | 2 | ||||
-rw-r--r-- | src/glsl/nir/nir_opt_peephole_select.c | 214 |
3 files changed, 217 insertions, 0 deletions
diff --git a/src/glsl/Makefile.sources b/src/glsl/Makefile.sources index 433224e4b24..38ffb850807 100644 --- a/src/glsl/Makefile.sources +++ b/src/glsl/Makefile.sources @@ -32,6 +32,7 @@ NIR_FILES = \ $(GLSL_SRCDIR)/nir/nir_opt_copy_propagate.c \ $(GLSL_SRCDIR)/nir/nir_opt_dce.c \ $(GLSL_SRCDIR)/nir/nir_opt_global_to_local.c \ + $(GLSL_SRCDIR)/nir/nir_opt_peephole_select.c \ $(GLSL_SRCDIR)/nir/nir_print.c \ $(GLSL_SRCDIR)/nir/nir_remove_dead_variables.c \ $(GLSL_SRCDIR)/nir/nir_to_ssa.c \ diff --git a/src/glsl/nir/nir.h b/src/glsl/nir/nir.h index 126e9c841c7..5c0a1e50ff1 100644 --- a/src/glsl/nir/nir.h +++ b/src/glsl/nir/nir.h @@ -1373,6 +1373,8 @@ bool nir_copy_prop(nir_shader *shader); bool nir_opt_dce_impl(nir_function_impl *impl); bool nir_opt_dce(nir_shader *shader); +bool nir_opt_peephole_select(nir_shader *shader); + #ifdef __cplusplus } /* extern "C" */ #endif diff --git a/src/glsl/nir/nir_opt_peephole_select.c b/src/glsl/nir/nir_opt_peephole_select.c new file mode 100644 index 00000000000..b3b54c1a1cb --- /dev/null +++ b/src/glsl/nir/nir_opt_peephole_select.c @@ -0,0 +1,214 @@ +/* + * Copyright © 2014 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + * Authors: + * Jason Ekstrand ([email protected]) + * + */ + +#include "nir.h" + +/* + * Implements a small peephole optimization that looks for + * + * if (cond) { + * <empty> + * } else { + * <empty> + * } + * phi + * ... + * phi + * + * and replaces it with a series of selects. It can also handle the case + * where, instead of being empty, the if may contain some move operations + * whose only use is one of the following phi nodes. This happens all the + * time when the SSA form comes from a conditional assignment with a + * swizzle. + */ + +struct peephole_select_state { + void *mem_ctx; + nir_function_impl *impl; + bool progress; +}; + +static bool +are_all_move_to_phi(nir_block *block) +{ + nir_foreach_instr(block, instr) { + if (instr->type != nir_instr_type_alu) + return false; + + /* It must be a move operation */ + nir_alu_instr *mov = nir_instr_as_alu(instr); + if (mov->op != nir_op_fmov && mov->op != nir_op_imov) + return false; + + /* Can't handle saturate */ + if (mov->dest.saturate) + return false; + + /* It must be SSA */ + if (!mov->dest.dest.is_ssa) + return false; + + /* It cannot have any if-uses */ + if (mov->dest.dest.ssa.if_uses->entries != 0) + return false; + + /* The only uses of this definition must be phi's in the successor */ + struct set_entry *entry; + set_foreach(mov->dest.dest.ssa.uses, entry) { + const nir_instr *dest_instr = entry->key; + if (dest_instr->type != nir_instr_type_phi || + dest_instr->block != block->successors[0]) + return false; + } + } + + return true; +} + +static bool +nir_opt_peephole_select_block(nir_block *block, void *void_state) +{ + struct peephole_select_state *state = void_state; + + /* If the block is empty, then it certainly doesn't have any phi nodes, + * so we can skip it. This also ensures that we do an early skip on the + * end block of the function which isn't actually attached to the CFG. + */ + if (exec_list_is_empty(&block->instr_list)) + return true; + + if (nir_cf_node_is_first(&block->cf_node)) + return true; + + nir_cf_node *prev_node = nir_cf_node_prev(&block->cf_node); + if (prev_node->type != nir_cf_node_if) + return true; + + nir_if *if_stmt = nir_cf_node_as_if(prev_node); + nir_cf_node *then_node = nir_if_first_then_node(if_stmt); + nir_cf_node *else_node = nir_if_first_else_node(if_stmt); + + /* We can only have one block in each side ... */ + if (nir_if_last_then_node(if_stmt) != then_node || + nir_if_last_else_node(if_stmt) != else_node) + return true; + + nir_block *then_block = nir_cf_node_as_block(then_node); + nir_block *else_block = nir_cf_node_as_block(else_node); + + /* ... and those blocks must only contain move-to-phi. */ + if (!are_all_move_to_phi(then_block) || !are_all_move_to_phi(else_block)) + return true; + + /* At this point, we know that the previous CFG node is an if-then + * statement containing only moves to phi nodes in this block. We can + * just remove that entire CF node and replace all of the phi nodes with + * selects. + */ + + nir_foreach_instr_safe(block, instr) { + if (instr->type != nir_instr_type_phi) + break; + + nir_phi_instr *phi = nir_instr_as_phi(instr); + nir_alu_instr *sel = nir_alu_instr_create(state->mem_ctx, nir_op_bcsel); + sel->src[0].src = nir_src_copy(if_stmt->condition, state->mem_ctx); + + assert(exec_list_length(&phi->srcs) == 2); + foreach_list_typed(nir_phi_src, src, node, &phi->srcs) { + assert(src->pred == then_block || src->pred == else_block); + assert(src->src.is_ssa); + + unsigned idx = src->pred == then_block ? 1 : 2; + + if (src->src.ssa->parent_instr->block == src->pred) { + /* We already know that this instruction must be a move with + * this phi's in this block as its only users. + */ + nir_alu_instr *mov = nir_instr_as_alu(src->src.ssa->parent_instr); + assert(mov->instr.type == nir_instr_type_alu); + assert(mov->op == nir_op_fmov || mov->op == nir_op_imov); + + sel->src[idx].src = nir_src_copy(mov->src[0].src, state->mem_ctx); + sel->src[idx].negate = mov->src[0].negate; + sel->src[idx].abs = mov->src[0].abs; + memcpy(sel->src[idx].swizzle, mov->src[0].swizzle, 4); + } else { + sel->src[idx].src = nir_src_copy(src->src, state->mem_ctx); + } + } + + sel->dest.dest.is_ssa = true; + nir_ssa_def_init(state->impl, &sel->instr, &sel->dest.dest.ssa, + phi->dest.ssa.num_components, phi->dest.ssa.name); + sel->dest.write_mask = (1 << phi->dest.ssa.num_components) - 1; + + nir_src sel_dest_src = { + .is_ssa = true, + .ssa = &sel->dest.dest.ssa, + }; + nir_ssa_def_rewrite_uses(&phi->dest.ssa, sel_dest_src, state->mem_ctx); + + nir_instr_insert_before(&phi->instr, &sel->instr); + nir_instr_remove(&phi->instr); + } + + nir_cf_node_remove(&if_stmt->cf_node); + state->progress = true; + + return true; +} + +static bool +nir_opt_peephole_select_impl(nir_function_impl *impl) +{ + struct peephole_select_state state; + + state.mem_ctx = ralloc_parent(impl); + state.progress = false; + state.impl = impl; + + nir_foreach_block(impl, nir_opt_peephole_select_block, &state); + + if (state.progress) + nir_metadata_dirty(impl, nir_metadata_none); + + return state.progress; +} + +bool +nir_opt_peephole_select(nir_shader *shader) +{ + bool progress = false; + + nir_foreach_overload(shader, overload) { + if (overload->impl) + progress |= nir_opt_peephole_select_impl(overload->impl); + } + + return progress; +} |