Welcome to ftp.nluug.nl Current directory: /ftp/os/BSD/NetBSD/NetBSD-release-10/xsrc/external/mit/MesaLib/dist/src/compiler/nir/ |
|
Contents of README:New IR, or NIR, is an IR for Mesa intended to sit below GLSL IR and Mesa IR. Its design inherits from the various IRs that Mesa has used in the past, as well as Direct3D assembly, and it includes a few new ideas as well. It is a flat (in terms of using instructions instead of expressions), typeless IR, similar to TGSI and Mesa IR. It also supports SSA (although it doesn't require it). Variables ========= NIR includes support for source-level GLSL variables through a structure mostly copied from GLSL IR. These will be used for linking and conversion from GLSL IR (and later, from an AST), but for the most part, they will be lowered to registers (see below) and loads/stores. Registers ========= Registers are light-weight; they consist of a structure that only contains its size, its index for liveness analysis, and an optional name for debugging. In addition, registers can be local to a function or global to the entire shader; the latter will be used in ARB_shader_subroutine for passing parameters and getting return values from subroutines. Registers can also be an array, in which case they can be accessed indirectly. Each ALU instruction (add, subtract, etc.) works directly with registers or SSA values (see below). SSA ======== Everywhere a register can be loaded/stored, an SSA value can be used instead. The only exception is that arrays/indirect addressing are not supported with SSA; although research has been done on extensions of SSA to arrays before, it's usually for the purpose of parallelization (which we're not interested in), and adds some overhead in the form of adding copies or extra arrays (which is much more expensive than introducing copies between non-array registers). SSA uses point directly to their corresponding definition, which in turn points to the instruction it is part of. This creates an implicit use-def chain and avoids the need for an external structure for each SSA register. Functions ========= Support for function calls is mostly similar to GLSL IR. Each shader contains a list of functions, and each function has a list of overloads. Each overload contains a list of parameters, and may contain an implementation which specifies the variables that correspond to the parameters and return value. Inlining a function, assuming it has a single return point, is as simple as copying its instructions, registers, and local variables into the target function and then inserting copies to and from the new parameters as appropriate. After functions are inlined and any non-subroutine functions are deleted, parameters and return variables will be converted to global variables and then global registers. We don't do this lowering earlier (i.e. the fortranizer idea) for a few reasons: - If we want to do optimizations before link time, we need to have the function signature available during link-time. - If we do any inlining before link time, then we might wind up with the inlined function and the non-inlined function using the same global variables/registers which would preclude optimization. Intrinsics ========= Any operation (other than function calls and textures) which touches a variable or is not referentially transparent is represented by an intrinsic. Intrinsics are similar to the idea of a "builtin function," i.e. a function declaration whose implementation is provided by the backend, except they are more powerful in the following ways: - They can also load and store registers when appropriate, which limits the number of variables needed in later stages of the IR while obviating the need for a separate load/store variable instruction. - Intrinsics can be marked as side-effect free, which permits them to be treated like any other instruction when it comes to optimizations. This allows load intrinsics to be represented as intrinsics while still being optimized away by dead code elimination, common subexpression elimination, etc. Intrinsics are used for: - Atomic operations - Memory barriers - Subroutine calls - Geometry shader emitVertex and endPrimitive - Loading and storing variables (before lowering) - Loading and storing uniforms, shader inputs and outputs, etc (after lowering) - Copying variables (cases where in GLSL the destination is a structure or array) - The kitchen sink - ... Textures ========= Unfortunately, there are far too many texture operations to represent each one of them with an intrinsic, so there's a special texture instruction similar to the GLSL IR one. The biggest difference is that, while the texture instruction has a sampler dereference field used just like in GLSL IR, this gets lowered to a texture unit index (with a possible indirect offset) while the type information of the original sampler is kept around for backends. Also, all the non-constant sources are stored in a single array to make it easier for optimization passes to iterate over all the sources. Control Flow ========= Like in GLSL IR, control flow consists of a tree of "control flow nodes", which include if statements and loops, and jump instructions (break, continue, and return). Unlike GLSL IR, though, the leaves of the tree aren't statements but basic blocks. Each basic block also keeps track of its successors and predecessors, and function implementations keep track of the beginning basic block (the first basic block of the function) and the ending basic block (a fake basic block that every return statement points to). Together, these elements make up the control flow graph, in this case a redundant piece of information on top of the control flow tree that will be used by almost all the optimizations. There are helper functions to add and remove control flow nodes that also update the control flow graph, and so usually it doesn't need to be touched by passes that modify control flow nodes. |
Name Last modified Size
Parent Directory - CVS/ 17-Dec-2022 21:58 - tests/ 16-Oct-2024 05:07 - README 10-Mar-2019 04:42 5.7K meson.build 09-May-2022 03:23 16K nir.c 09-May-2022 03:23 79K nir.h 09-May-2022 03:23 179K nir_algebraic.py 09-May-2022 03:23 47K nir_builder.h 09-May-2022 03:23 57K nir_builder_opcodes_h.py 09-May-2022 03:23 5.4K nir_builtin_builder.c 09-May-2022 03:23 15K nir_builtin_builder.h 09-May-2022 03:27 7.6K nir_clone.c 09-May-2022 03:23 24K nir_constant_expressions.h 09-May-2022 03:23 1.6K nir_constant_expressions.py 09-May-2022 03:23 16K nir_control_flow.c 09-May-2022 03:23 24K nir_control_flow.h 09-May-2022 03:23 6.9K nir_control_flow_private.h 10-Mar-2019 04:42 1.5K nir_conversion_builder.h 09-May-2022 03:23 18K nir_convert_ycbcr.c 09-May-2022 03:23 5.6K nir_deref.c 09-May-2022 03:23 45K nir_deref.h 09-May-2022 03:23 3.0K nir_divergence_analysis.c 09-May-2022 03:23 36K nir_dominance.c 09-May-2022 03:23 9.3K nir_format_convert.h 09-May-2022 03:23 16K nir_from_ssa.c 09-May-2022 03:23 37K nir_gather_info.c 09-May-2022 03:23 35K nir_gather_ssa_types.c 09-May-2022 03:23 8.6K nir_gather_xfb_info.c 09-May-2022 03:23 10K nir_gs_count_vertices.c 09-May-2022 03:23 4.2K nir_inline_functions.c 09-May-2022 03:23 11K nir_inline_helpers.h 09-May-2022 03:23 4.7K nir_inline_uniforms.c 09-May-2022 03:23 16K nir_instr_set.c 09-May-2022 03:23 25K nir_instr_set.h 09-May-2022 03:23 2.8K nir_intrinsics.py 09-May-2022 03:23 58K nir_intrinsics_c.py 09-May-2022 03:23 2.9K nir_intrinsics_h.py 09-May-2022 03:23 2.2K nir_intrinsics_indices_h.py 09-May-2022 03:23 3.1K nir_linking_helpers.c 09-May-2022 03:23 56K nir_liveness.c 09-May-2022 03:23 14K nir_loop_analyze.c 09-May-2022 03:23 45K nir_loop_analyze.h 09-May-2022 03:23 3.5K nir_lower_alpha_test.c 09-May-2022 03:23 4.7K nir_lower_alu.c 09-May-2022 03:23 10K nir_lower_alu_to_scalar.c 09-May-2022 03:23 13K nir_lower_amul.c 09-May-2022 03:23 10K nir_lower_array_deref_of_vec.c 09-May-2022 03:23 7.5K nir_lower_atomics_to_ssbo.c 09-May-2022 03:23 8.6K nir_lower_bit_size.c 09-May-2022 03:23 13K nir_lower_bitmap.c 09-May-2022 03:23 4.8K nir_lower_blend.c 09-May-2022 03:23 12K nir_lower_blend.h 09-May-2022 03:23 2.1K nir_lower_bool_to_bitsize.c 09-May-2022 03:23 14K nir_lower_bool_to_float.c 09-May-2022 03:23 6.4K nir_lower_bool_to_int32.c 09-May-2022 03:23 6.1K nir_lower_clamp_color_outputs.c 09-May-2022 03:23 3.9K nir_lower_clip.c 09-May-2022 03:23 16K nir_lower_clip_cull_distance_arrays.c 09-May-2022 03:23 4.7K nir_lower_clip_disable.c 09-May-2022 03:23 5.5K nir_lower_clip_halfz.c 09-May-2022 03:23 2.9K nir_lower_convert_alu_types.c 09-May-2022 03:23 6.3K nir_lower_discard_or_demote.c 09-May-2022 03:23 7.8K nir_lower_double_ops.c 09-May-2022 03:23 25K nir_lower_drawpixels.c 09-May-2022 03:23 10K nir_lower_fb_read.c 09-May-2022 03:23 4.3K nir_lower_flatshade.c 09-May-2022 03:23 1.7K nir_lower_flrp.c 09-May-2022 03:23 21K nir_lower_fp16_conv.c 09-May-2022 03:23 9.6K nir_lower_fragcolor.c 09-May-2022 03:23 4.2K nir_lower_fragcoord_wtrans.c 09-May-2022 03:23 2.7K nir_lower_frexp.c 09-May-2022 03:23 7.3K nir_lower_global_vars_to_local.c 09-May-2022 03:23 3.7K nir_lower_goto_ifs.c 09-May-2022 03:23 33K nir_lower_gs_intrinsics.c 09-May-2022 03:23 15K nir_lower_idiv.c 09-May-2022 03:23 9.4K nir_lower_image.c 09-May-2022 03:23 3.4K nir_lower_indirect_derefs.c 09-May-2022 03:23 8.8K nir_lower_input_attachments.c 09-May-2022 03:23 7.5K nir_lower_int64.c 09-May-2022 03:23 43K nir_lower_int_to_float.c 09-May-2022 03:23 6.3K nir_lower_interpolation.c 09-May-2022 03:23 5.3K nir_lower_io.c 09-May-2022 03:23 96K nir_lower_io_arrays_to_elements.c 09-May-2022 03:23 15K nir_lower_io_to_scalar.c 09-May-2022 03:23 14K nir_lower_io_to_temporaries.c 09-May-2022 03:23 14K nir_lower_io_to_vector.c 09-May-2022 03:23 23K nir_lower_is_helper_invocation.c 09-May-2022 03:23 4.4K nir_lower_load_const_to_scalar.c 09-May-2022 03:23 3.4K nir_lower_locals_to_regs.c 09-May-2022 03:23 8.9K nir_lower_mediump.c 09-May-2022 03:23 22K nir_lower_memcpy.c 09-May-2022 03:23 7.1K nir_lower_memory_model.c 09-May-2022 03:23 9.5K nir_lower_multiview.c 09-May-2022 03:23 12K nir_lower_non_uniform_access.c 09-May-2022 03:23 13K nir_lower_packing.c 09-May-2022 03:23 5.0K nir_lower_passthrough_edgeflags.c 09-May-2022 03:23 3.7K nir_lower_patch_vertices.c 09-May-2022 03:23 3.9K nir_lower_phis_to_scalar.c 09-May-2022 03:23 11K nir_lower_pntc_ytransform.c 09-May-2022 03:23 4.9K nir_lower_point_size.c 09-May-2022 03:23 3.7K nir_lower_point_size_mov.c 09-May-2022 03:23 2.9K nir_lower_printf.c 09-May-2022 03:23 5.9K nir_lower_readonly_images_to_tex.c 09-May-2022 03:23 7.4K nir_lower_regs_to_ssa.c 09-May-2022 03:23 9.8K nir_lower_returns.c 09-May-2022 03:23 10K nir_lower_samplers.c 09-May-2022 03:23 5.7K nir_lower_scratch.c 09-May-2022 03:23 7.7K nir_lower_shader_calls.c 09-May-2022 03:23 40K nir_lower_ssbo.c 09-May-2022 03:23 7.0K nir_lower_subgroups.c 09-May-2022 03:23 29K nir_lower_system_values.c 09-May-2022 03:23 20K nir_lower_sysvals_to_varyings.c 09-May-2022 03:23 2.5K nir_lower_tex.c 09-May-2022 03:23 51K nir_lower_texcoord_replace.c 09-May-2022 03:23 5.5K nir_lower_to_source_mods.c 09-May-2022 03:23 7.1K nir_lower_two_sided_color.c 09-May-2022 03:23 6.9K nir_lower_ubo_vec4.c 09-May-2022 03:23 7.6K nir_lower_undef_to_zero.c 09-May-2022 03:23 2.5K nir_lower_uniforms_to_ubo.c 09-May-2022 03:23 6.9K nir_lower_var_copies.c 09-May-2022 03:23 5.9K nir_lower_variable_initializers.c 09-May-2022 03:23 7.1K nir_lower_vars_to_ssa.c 09-May-2022 03:23 26K nir_lower_vec3_to_vec4.c 09-May-2022 03:23 5.5K nir_lower_vec_to_movs.c 09-May-2022 03:23 10K nir_lower_viewport_transform.c 09-May-2022 03:23 4.3K nir_lower_wpos_center.c 09-May-2022 03:23 3.9K nir_lower_wpos_ytransform.c 09-May-2022 03:23 13K nir_lower_wrmasks.c 09-May-2022 03:23 7.5K nir_metadata.c 09-May-2022 03:23 3.3K nir_move_vec_src_uses_to_dest.c 09-May-2022 03:23 6.7K nir_normalize_cubemap_coords.c 10-Mar-2019 04:42 3.8K nir_opcodes.py 09-May-2022 03:23 55K nir_opcodes_c.py 09-May-2022 03:23 4.9K nir_opcodes_h.py 09-May-2022 03:23 1.6K nir_opt_access.c 09-May-2022 03:23 13K nir_opt_algebraic.py 09-May-2022 03:23 146K nir_opt_barriers.c 09-May-2022 03:23 2.7K nir_opt_combine_stores.c 09-May-2022 03:23 16K nir_opt_comparison_pre.c 09-May-2022 03:23 13K nir_opt_conditional_discard.c 09-May-2022 03:23 4.7K nir_opt_constant_folding.c 09-May-2022 03:23 12K nir_opt_copy_prop_vars.c 09-May-2022 03:23 45K nir_opt_copy_propagate.c 09-May-2022 03:23 5.4K nir_opt_cse.c 09-May-2022 03:23 2.3K nir_opt_dce.c 09-May-2022 03:23 7.9K nir_opt_dead_cf.c 09-May-2022 03:23 14K nir_opt_dead_write_vars.c 09-May-2022 03:23 9.7K nir_opt_find_array_copies.c 09-May-2022 03:23 22K nir_opt_fragdepth.c 09-May-2022 03:23 3.8K nir_opt_gcm.c 09-May-2022 03:23 27K nir_opt_idiv_const.c 09-May-2022 03:23 8.7K nir_opt_if.c 09-May-2022 03:23 51K nir_opt_intrinsics.c 09-May-2022 03:23 7.0K nir_opt_large_constants.c 09-May-2022 03:23 14K nir_opt_load_store_vectorize.c 09-May-2022 03:23 49K nir_opt_loop_unroll.c 09-May-2022 03:23 38K nir_opt_memcpy.c 09-May-2022 03:23 6.7K nir_opt_move.c 09-May-2022 03:23 6.0K nir_opt_move_discards_to_top.c 09-May-2022 03:23 8.3K nir_opt_offsets.c 09-May-2022 03:23 5.2K nir_opt_peephole_select.c 09-May-2022 03:23 17K nir_opt_phi_precision.c 09-May-2022 03:23 14K nir_opt_rematerialize_compares.c 09-May-2022 03:23 5.9K nir_opt_remove_phis.c 09-May-2022 03:23 5.8K nir_opt_shrink_vectors.c 09-May-2022 03:23 9.8K nir_opt_sink.c 09-May-2022 03:23 7.6K nir_opt_trivial_continues.c 09-May-2022 03:23 4.8K nir_opt_undef.c 09-May-2022 03:23 5.7K nir_opt_uniform_atomics.c 09-May-2022 03:23 11K nir_opt_vectorize.c 09-May-2022 03:23 14K nir_phi_builder.c 09-May-2022 03:23 11K nir_phi_builder.h 10-Mar-2019 04:42 4.7K nir_print.c 09-May-2022 03:23 49K nir_propagate_invariant.c 09-May-2022 03:23 6.7K nir_range_analysis.c 09-May-2022 03:23 68K nir_range_analysis.h 09-May-2022 03:23 1.9K nir_remove_dead_variables.c 09-May-2022 03:23 7.0K nir_repair_ssa.c 09-May-2022 03:23 6.6K nir_schedule.c 09-May-2022 03:23 35K nir_schedule.h 09-May-2022 03:23 3.1K nir_search.c 09-May-2022 03:23 31K nir_search.h 09-May-2022 03:23 6.8K nir_search_helpers.h 09-May-2022 03:23 17K nir_serialize.c 09-May-2022 03:23 64K nir_serialize.h 09-May-2022 03:23 1.6K nir_split_per_member_structs.c 09-May-2022 03:23 6.5K nir_split_var_copies.c 09-May-2022 03:23 5.5K nir_split_vars.c 09-May-2022 03:23 58K nir_sweep.c 09-May-2022 03:23 4.8K nir_to_lcssa.c 09-May-2022 03:23 13K nir_validate.c 09-May-2022 03:23 61K nir_vla.h 10-Mar-2019 04:42 2.1K nir_vulkan.h 09-May-2022 03:23 374 nir_worklist.c 09-May-2022 03:23 3.8K nir_worklist.h 09-May-2022 03:23 4.8K nir_xfb_info.h 24-Sep-2019 18:46 2.4K
NLUUG - Open Systems. Open Standards
Become a member
and get discounts on conferences and more, see the NLUUG website!