Searched refs:parallelcopies (Results 1 – 4 of 4) sorted by relevance
| /external/mesa3d/src/amd/compiler/ |
| D | aco_register_allocation.cpp | 551 std::vector<std::pair<Operand, Definition>>& parallelcopies, in update_renames() argument
555 for (std::pair<Operand, Definition>& copy : parallelcopies) { in update_renames()
561 for (std::pair<Operand, Definition>& other : parallelcopies) { in update_renames()
583 for (std::pair<Operand, Definition>& pc : parallelcopies) { in update_renames()
767 std::vector<std::pair<Operand, Definition>>& parallelcopies, in get_regs_for_copies() argument
833 parallelcopies.emplace_back(pc_op, pc_def); in get_regs_for_copies()
915 …if (!get_regs_for_copies(ctx, reg_file, parallelcopies, new_vars, lb, ub, instr, def_reg_lo, def_r… in get_regs_for_copies()
925 parallelcopies.emplace_back(pc_op, pc_def); in get_regs_for_copies()
934 std::vector<std::pair<Operand, Definition>>& parallelcopies, in get_reg_impl() argument
1106 parallelcopies.insert(parallelcopies.end(), pc.begin(), pc.end()); in get_reg_impl()
[all …]
|
| D | README.md | 93 The next step is a pass out of SSA by inserting parallelcopies at the end of blocks to match the ph…
|
| /external/mesa3d/docs/relnotes/ |
| D | 20.0.0.rst | 767 - aco: handle phi affinities transitively through parallelcopies
768 - aco: ignore parallelcopies to the same register on jump threading
|
| D | 20.2.0.rst | 1234 - aco: coalesce parallelcopies during register allocation
|