summaryrefslogtreecommitdiffstats
path: root/src/compiler/nir
Commit message (Collapse)AuthorAgeFilesLines
* nir/algebraic: Don't put quotes around floating point literalsIan Romanick2018-12-182-5/+13
| | | | | | | | | | | | | | | | | | The quotation marks around 1.0 cause it to be treated as a string instead of a floating point value. The generator then treats it as an arbitrary variable replacement, so any iand involving a ('ineg', ('b2i', a)) matches. v2: Remove misleading comment about sized literals (suggested by Timothy). Add assertion that the name of a varible is entierly alphabetic (suggested by Jason). Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Tested-by: Timothy Arceri <[email protected]> [v1] Reviewed-by: Timothy Arceri <[email protected]> [v1] Fixes: 6bcd2af0863 ("nir/algebraic: Add some optimizations for D3D-style Booleans") Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=109075
* nir: Add a new lowering option to lower 3D surfaces from txd to txl.Sagar Ghuge2018-12-182-1/+8
| | | | | | | | | Tested on gen9. v2: Rename lower_txd_3d_surafaces flag to lower_txd_3d (Jason Ekstrand) Signed-off-by: Sagar Ghuge <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir/dead_write_vars: Get modes directly from derefsJason Ekstrand2018-12-181-2/+1
| | | | | | | | | Instead of going all the way back to the variable, just look at the deref. The modes are guaranteed to be the same by nir_validate whenever the variable can be found. This fixes clear_unused_for_modes for derefs that don't have an accessible variable. Reviewed-by: Timothy Arceri <[email protected]>
* nir/copy_prop_vars: Get modes directly from derefsJason Ekstrand2018-12-181-6/+2
| | | | | | | | | Instead of going all the way back to the variable, just look at the deref. The modes are guaranteed to be the same by nir_validate whenever the variable can be found. This fixes apply_barrier_for_modes for derefs that don't have an accessible variable. Reviewed-by: Timothy Arceri <[email protected]>
* nir/lower_wpos_center: Look at derefs for modesJason Ekstrand2018-12-181-2/+4
| | | | | | | | | This is instead of looking all the way back to the variable which may not exist for all derefs. This makes this code properly ignore casts with modes other than the mode[s] we care about (where casts aren't allowed). Reviewed-by: Timothy Arceri <[email protected]>
* nir/lower_io_to_scalar: Look at derefs for modesJason Ekstrand2018-12-181-3/+6
| | | | | | | | | This is instead of looking all the way back to the variable which may not exist for all derefs. This makes this code properly ignore casts with modes other than the mode[s] we care about (where casts aren't allowed). Reviewed-by: Timothy Arceri <[email protected]>
* nir/lower_io_arrays_to_elements: Look at derefs for modesJason Ekstrand2018-12-181-5/+8
| | | | | | | | | This is instead of looking all the way back to the variable which may not exist for all derefs. This makes this code properly ignore casts with modes other than the mode[s] we care about (where casts aren't allowed). Reviewed-by: Timothy Arceri <[email protected]>
* nir/linking_helpers: Look at derefs for modesJason Ekstrand2018-12-181-12/+11
| | | | | | | | | This is instead of looking all the way back to the variable which may not exist for all derefs. This makes this code properly ignore casts with modes other than the mode[s] we care about (where casts aren't allowed). Reviewed-by: Timothy Arceri <[email protected]>
* nir/propagate_invariant: Skip unknown varsJason Ekstrand2018-12-181-1/+1
| | | | | | | | If we can't find the variable from the deref, just assume it isn't invariant and continue on. This can happen if, for instance, we're writing to a deref that points into an SSBO. Reviewed-by: Timothy Arceri <[email protected]>
* Revert "nir/lower_indirect: Bail early if modes == 0"Ian Romanick2018-12-181-3/+0
| | | | | | | | | | | | | | | | | "There's no point in walking the program if we're never going to actually lower anything." Except we might lower compacted local arrays. In that case, modes will be 0, but there is still lowering to be done. This reverts commit 7f75cf2a9408b9af562e033ef6c1d1fd15141421. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=109081 Suggested-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Lionel Landwerlin <[email protected]> Tested-by: Clayton Craft <[email protected]> Cc: Kenneth Graunke <[email protected]>
* nir/opt_peephole_select: Don't peephole_select expensive math instructionsIan Romanick2018-12-172-9/+32
| | | | | | | | | | | | | | | | On some GPUs, especially older Intel GPUs, some math instructions are very expensive. On those architectures, don't reduce flow control to a csel if one of the branches contains one of these expensive math instructions. This prevents a bunch of cycle count regressions on pre-Gen6 platforms with a later patch (intel/compiler: More peephole select for pre-Gen6). v2: Remove stray #if block. Noticed by Thomas. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Thomas Helland <[email protected]> Reviewed-by: Lionel Landwerlin <[email protected]>
* nir/opt_peephole_select: Don't try to remove flow control around indirect loadsIan Romanick2018-12-172-11/+29
| | | | | | | | | | | | | | | | | | | That flow control may be trying to avoid invalid loads. On at least some platforms, those loads can also be expensive. No shader-db changes on any Intel platform (even with the later patch "intel/compiler: More peephole select"). v2: Add a 'indirect_load_ok' flag to nir_opt_peephole_select. Suggested by Rob. See also the big comment in src/intel/compiler/brw_nir.c. v3: Use nir_deref_instr_has_indirect instead of deref_has_indirect (from nir_lower_io_arrays_to_elements.c). v4: Fix inverted condition in brw_nir.c. Noticed by Lionel. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Lionel Landwerlin <[email protected]>
* nir: Fix clamping of uints for image store lowering.Eric Anholt2018-12-171-1/+1
| | | | | | | | | | I botched some copy-and-paste and clamped to signed int max instead of uint max. Fixes KHR-GL46.shader_image_load_store.multiple-uniforms on skl. Fixes: d3e046e76c06 ("nir: Pull some of intel's image load/store format conversion to nir_format.h") Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Release per-block metadata in nir_sweepIan Romanick2018-12-161-0/+9
| | | | | | | | | | | | | | | nir_sweep already marks all metadata invalid, so it is safe to release the memory here too. mean soft fp64 using uint64: 1,342,759,331 => 1,010,670,475 gfxbench5 aztec ruins high 11: 63,555,571 => 61,889,811 deus ex mankind divided 148: 62,845,304 => 62,829,640 deus ex mankind divided 2890: 71,922,686 => 71,922,686 dirt showdown 676: 69,238,607 => 69,238,607 dolphin ubershaders 210: 77,822,072 => 77,822,072 Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Fix holes in nir_instrIan Romanick2018-12-161-5/+5
| | | | | | | | | | | | | | | | Found using pahole. Changes in peak memory usage according to Valgrind massif: mean soft fp64 using uint64: 1,343,991,403 => 1,342,759,331 gfxbench5 aztec ruins high 11: 63,619,971 => 63,555,571 deus ex mankind divided 148: 62,887,728 => 62,845,304 deus ex mankind divided 2890: 72,399,750 => 71,922,686 dirt showdown 676: 69,464,023 => 69,238,607 dolphin ubershaders 210: 78,359,728 => 77,822,072 Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir/phi_builder: Use per-value hash table to store [block] -> def mappingIan Romanick2018-12-161-9/+36
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Replace the old array in each value with a hash table in each value. Changes in peak memory usage according to Valgrind massif: mean soft fp64 using uint64: 5,499,875,082 => 1,343,991,403 gfxbench5 aztec ruins high 11: 63,619,971 => 63,619,971 deus ex mankind divided 148: 62,887,728 => 62,887,728 deus ex mankind divided 2890: 72,402,222 => 72,399,750 dirt showdown 676: 74,466,431 => 69,464,023 dolphin ubershaders 210: 109,630,376 => 78,359,728 Run-time change for a full run on shader-db on my Haswell desktop (with -march=native) is 1.22245% +/- 0.463879% (n=11). This is about +2.9 seconds on a 237 second run. The first time I sent this version of this patch out, the run-time data was quite different. I had misconfigured the script that ran the test, and none of the tests from higher GLSL versions were run. These are generally more complex shaders, and they are more affected by this change. The previous version of this patch used a single hash table for the whole phi builder. The mapping was from [value, block] -> def, so a separate allocation was needed for each [value, block] tuple. There was quite a bit of per-allocation overhead (due to ralloc), so the patch was followed by a patch that added the use of the slab allocator. The results of those two patches was not quite as good: mean soft fp64 using uint64: 5,499,875,082 => 1,343,991,403 gfxbench5 aztec ruins high 11: 63,619,971 => 63,619,971 deus ex mankind divided 148: 62,887,728 => 62,887,728 deus ex mankind divided 2890: 72,402,222 => 72,402,222 * dirt showdown 676: 74,466,431 => 72,443,591 * dolphin ubershaders 210: 109,630,376 => 81,034,320 * The * denote tests that are better now. In the tests that are the same in both patches, the "after" peak memory usage was at a different location. I did not check the local peaks. Signed-off-by: Ian Romanick <[email protected]> Suggested-by: Jason Ekstrand <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir/algebraic: Add some optimizations for D3D-style BooleansJason Ekstrand2018-12-161-0/+13
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | D3D Booleans use a 32-bit 0/-1 representation. Because this previously matched NIR exactly, we didn't have to really optimize for it. Now that we have 1-bit Booleans, we need some specific optimizations to chew through the D3D12-style Booleans. Shader-db results on Kaby Lake: total instructions in shared programs: 15136811 -> 14967944 (-1.12%) instructions in affected programs: 2457021 -> 2288154 (-6.87%) helped: 8318 HURT: 10 total cycles in shared programs: 373544524 -> 359701825 (-3.71%) cycles in affected programs: 151029683 -> 137186984 (-9.17%) helped: 7749 HURT: 682 total loops in shared programs: 4431 -> 4399 (-0.72%) loops in affected programs: 32 -> 0 helped: 21 HURT: 0 total spills in shared programs: 10290 -> 10051 (-2.32%) spills in affected programs: 2532 -> 2293 (-9.44%) helped: 18 HURT: 18 total fills in shared programs: 22203 -> 21732 (-2.12%) fills in affected programs: 3319 -> 2848 (-14.19%) helped: 18 HURT: 18 Note that a large chunk of the improvement fixing regressions caused by switching to 1-bit Booleans. Previously, our ability to optimize D3D booleans was improved by using the D3D representation directly in NIR. Now that NIR does 1-bit bools, we need a few more optimizations. Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Eric Anholt <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir/algebraic: Optimize 1-bit BooleansJason Ekstrand2018-12-162-86/+57
| | | | | | Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir: Switch to using 1-bit Booleans for almost everythingJason Ekstrand2018-12-169-74/+50
| | | | | | | | | | | | | | | | This is a squash of a few distinct changes: glsl,spirv: Generate 1-bit Booleans Revert "Use 32-bit opcodes in the NIR producers and optimizations" Revert "nir/builder: Generate 32-bit bool opcodes transparently" nir/builder: Generate 1-bit Booleans in nir_build_imm_bool Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir: Add a bool to int32 lowering passJason Ekstrand2018-12-163-0/+162
| | | | | | | | We also enable it in all of the NIR drivers. Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir: Add 1-bit Boolean opcodesJason Ekstrand2018-12-163-3/+34
| | | | | | | | | We also have to add support for 1-bit integers while we're here so we get 1-bit variants of iand, ior, and inot. Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir/algebraic: Generalize an optimizationJason Ekstrand2018-12-161-2/+9
| | | | | | | | This just makes it nicely scale across bit sizes. Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir/large_constants: Properly handle 1-bit boolsJason Ekstrand2018-12-161-1/+23
| | | | | | Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir: Add support for 1-bit data typesJason Ekstrand2018-12-1610-24/+83
| | | | | | | | | | | | | | This commit adds support for 1-bit Booleans and integers. Booleans obviously take a value of true or false. Because we have to define the semantics of 1-bit signed and unsigned integers, we define uint1_t to take values of 0 and 1 and int1_t to take values of 0 and -1. 1-bit arithmetic is then well-defined in the usual way, just with fewer bits. The definition of int1_t and uint1_t doesn't usually matter but we do need something for purposes of constant folding. Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir/constant_expressions: Rework Boolean handlingJason Ekstrand2018-12-161-20/+12
| | | | | | | | | | | | | | This commit contains three related changes. First, we define boolN_t for N = 8, 16, and 64 and move the definition of boolN_vec to the loop with the other vec definitions. Second, there's no reason why we need the != 0 on the source because that happens implicitly when it's converted to bool. Third, for destinations, we use a signed integer type and just do -(int)bool_val which will give us the 0/-1 behavior we want and neatly scales to all bit widths. Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir: Rename Boolean-related opcodes to include 32 in the nameJason Ekstrand2018-12-169-49/+116
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This is a squash of a bunch of individual changes: nir/builder: Generate 32-bit bool opcodes transparently nir/algebraic: Remap Boolean opcodes to the 32-bit variant Use 32-bit opcodes in the NIR producers and optimizations Generated with a little hand-editing and the following sed commands: sed -i 's/nir_op_ball_fequal/nir_op_b32all_fequal/g' **/*.c sed -i 's/nir_op_bany_fnequal/nir_op_b32any_fnequal/g' **/*.c sed -i 's/nir_op_ball_iequal/nir_op_b32all_iequal/g' **/*.c sed -i 's/nir_op_bany_inequal/nir_op_b32any_inequal/g' **/*.c sed -i 's/nir_op_\([fiu]lt\)/nir_op_\132/g' **/*.c sed -i 's/nir_op_\([fiu]ge\)/nir_op_\132/g' **/*.c sed -i 's/nir_op_\([fiu]ne\)/nir_op_\132/g' **/*.c sed -i 's/nir_op_\([fiu]eq\)/nir_op_\132/g' **/*.c sed -i 's/nir_op_\([fi]\)ne32g/nir_op_\1neg/g' **/*.c sed -i 's/nir_op_bcsel/nir_op_b32csel/g' **/*.c Use 32-bit opcodes in the NIR back-ends Generated with a little hand-editing and the following sed commands: sed -i 's/nir_op_ball_fequal/nir_op_b32all_fequal/g' **/*.c sed -i 's/nir_op_bany_fnequal/nir_op_b32any_fnequal/g' **/*.c sed -i 's/nir_op_ball_iequal/nir_op_b32all_iequal/g' **/*.c sed -i 's/nir_op_bany_inequal/nir_op_b32any_inequal/g' **/*.c sed -i 's/nir_op_\([fiu]lt\)/nir_op_\132/g' **/*.c sed -i 's/nir_op_\([fiu]ge\)/nir_op_\132/g' **/*.c sed -i 's/nir_op_\([fiu]ne\)/nir_op_\132/g' **/*.c sed -i 's/nir_op_\([fiu]eq\)/nir_op_\132/g' **/*.c sed -i 's/nir_op_\([fi]\)ne32g/nir_op_\1neg/g' **/*.c sed -i 's/nir_op_bcsel/nir_op_b32csel/g' **/*.c Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir/algebraic: Make an optimization more specificJason Ekstrand2018-12-161-1/+1
| | | | | | | | Later in this series, bool is not going to imply 32-bit. Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir: Drop support for lower_b2fJason Ekstrand2018-12-162-7/+1
| | | | | | | | | | This was originally added for the out-of-tree Mali driver but I think we've all agreed it's easy enough for them to just do in their back-end. Reviewed-by: Alyssa Rosenzweig <[email protected]> Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir/algebraic: Optimize x2b(xneg(a)) -> aJason Ekstrand2018-12-161-0/+2
| | | | | | | | | | | | | | | Shader-db results on Kaby Lake: total instructions in shared programs: 15072525 -> 15072525 (0.00%) instructions in affected programs: 0 -> 0 helped: 0 HURT: 0 This helps prevent regressions in later commits. Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir/constant_folding: Fix source bit size logicJason Ekstrand2018-12-161-1/+2
| | | | | | | | | | | | | Instead of looking at input_sizes[i] which contains the number of components for each source, we look at the bit size of input_types[i]. This fixes a regression in the 1-bit boolean series though I have no idea how we haven't seen it before now. Fixes: 35baee5dce5 "nir/constant_folding: fix incorrect bit-size check" Fixes: 9076c4e289d "nir: update opcode definitions for different bit sizes" Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]> Tested-by: Bas Nieuwenhuizen <[email protected]>
* nir/lower_idiv: Use ilt instead of bit twiddlingJason Ekstrand2018-12-161-1/+1
| | | | | | | | | The previous code was creating a boolean by doing an arithmetic right- shift by 31 which produces a boolean which is true if the argument is negative. This is the same as the expression r < 0 which is much simpler and doesn't depend on NIR's representation of booleans. Reviewed-by: Eric Anholt <[email protected]>
* nir: fix constness in nir_intrinsic_align()Rhys Perry2018-12-161-1/+1
| | | | | Signed-off-by: Rhys Perry <[email protected]> Reviewed-by: Samuel Pitoiset <[email protected]>
* nir/phi_builder: Internal users should use ↵Ian Romanick2018-12-141-2/+2
| | | | | | | nir_phi_builder_value_set_block_def too Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir: fix opt_if_loop_last_continue()Timothy Arceri2018-12-141-2/+6
| | | | | | | | | | | | | | | | | | | | | | | The pass did not correctly handle loops ending in: if ssa_7 { block block_8: /* preds: block_7 */ continue /* succs: block_1 */ } else { block block_9: /* preds: block_7 */ break /* succs: block_11 */ } The break will get eliminated by another opt but if this pass gets called first (as it does on RADV) we ended up inserting instructions after the break. Fixes: 5921a19d4b0c ("nir: add if opt opt_if_loop_last_continue()") Reviewed-by: Dave Airlie <[email protected]>
* nir: Move intel's half-float image store lowering to to nir_format.h.Eric Anholt2018-12-131-0/+13
| | | | | | | | I needed the same function for v3d. This was originally in d3e046e76c06 ("nir: Pull some of intel's image load/store format conversion to nir_format.h") before we made am istake about simplifying the function. Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Print the format of image variables.Eric Anholt2018-12-131-0/+47
| | | | | | | | This helps a lot when debugging image load/store lowering on large testcases. Unfortunately the Mesa enum name stuff is under src/mesa and we can't get at it from the compiler. Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Add a pass for lowering integer division by constantsJason Ekstrand2018-12-133-0/+218
| | | | | | | | | | | It's a reasonably well-known fact in the world of compilers that integer divisions by constants can be replaced by a multiply, an add, and some shifts. This commit adds such an optimization to NIR for easiest case of udiv. Other division operations will be added in following commits. In order to provide some additional driver control, the pass takes a minimum bit size to optimize. Reviewed-by: Ian Romanick [email protected]
* nir: Add a saturated unsigned integer add opcodeIan Romanick2018-12-131-0/+2
| | | | Reviewed-by: Jason Ekstrand <[email protected]>
* nir/lower_int64: Add support for [iu]mul_highJason Ekstrand2018-12-132-0/+67
| | | | Reviewed-by: Ian Romanick [email protected]
* nir: Allow [iu]mul_high on non-32-bit typesJason Ekstrand2018-12-132-4/+40
| | | | Reviewed-by: Ian Romanick [email protected]
* nir: remove unused variableAlejandro Piñeiro2018-12-131-1/+0
| | | | | | | To avoid the following warning: ./src/compiler/nir/nir_loop_analyze.c:807:16: warning: unused variable ‘ns’ [-Wunused-variable] nir_shader *ns = impl->function->shader; Reviewed-by: Lionel Landwerlin <[email protected]>
* nir: Pull some of intel's image load/store format conversion to nir_format.hEric Anholt2018-12-121-0/+38
| | | | | | | | | I needed the same functions for v3d. Note that the color value in the Intel lowering has already been cut down to image.chans num_components. v2: Drop the half float one, since it was a 1-liner after cleanup. Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Add some more consts to the nir_format_convert.h helpers.Eric Anholt2018-12-121-7/+6
| | | | | | | Most of the bits were constant, but a few were missed. Avoids warnings from v3d's upcoming static const bits declarations. Reviewed-by: Jason Ekstrand <[email protected]>
* nir: detect more induction variablesTimothy Arceri2018-12-131-0/+36
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This allows loop analysis to detect inductions variables that are incremented in both branches of an if rather than in a main loop block. For example: loop { block block_1: /* preds: block_0 block_7 */ vec1 32 ssa_8 = phi block_0: ssa_4, block_7: ssa_20 vec1 32 ssa_9 = phi block_0: ssa_0, block_7: ssa_4 vec1 32 ssa_10 = phi block_0: ssa_1, block_7: ssa_4 vec1 32 ssa_11 = phi block_0: ssa_2, block_7: ssa_21 vec1 32 ssa_12 = phi block_0: ssa_3, block_7: ssa_22 vec4 32 ssa_13 = vec4 ssa_12, ssa_11, ssa_10, ssa_9 vec1 32 ssa_14 = ige ssa_8, ssa_5 /* succs: block_2 block_3 */ if ssa_14 { block block_2: /* preds: block_1 */ break /* succs: block_8 */ } else { block block_3: /* preds: block_1 */ /* succs: block_4 */ } block block_4: /* preds: block_3 */ vec1 32 ssa_15 = ilt ssa_6, ssa_8 /* succs: block_5 block_6 */ if ssa_15 { block block_5: /* preds: block_4 */ vec1 32 ssa_16 = iadd ssa_8, ssa_7 vec1 32 ssa_17 = load_const (0x3f800000 /* 1.000000*/) /* succs: block_7 */ } else { block block_6: /* preds: block_4 */ vec1 32 ssa_18 = iadd ssa_8, ssa_7 vec1 32 ssa_19 = load_const (0x3f800000 /* 1.000000*/) /* succs: block_7 */ } block block_7: /* preds: block_5 block_6 */ vec1 32 ssa_20 = phi block_5: ssa_16, block_6: ssa_18 vec1 32 ssa_21 = phi block_5: ssa_17, block_6: ssa_4 vec1 32 ssa_22 = phi block_5: ssa_4, block_6: ssa_19 /* succs: block_1 */ } Unfortunatly GCM could move the addition out of the if for us (making this patch unrequired) but we still cannot enable the GCM pass without regressions. This unrolls a loop in Rise of The Tomb Raider. vkpipeline-db results (VEGA): Totals from affected shaders: SGPRS: 88 -> 96 (9.09 %) VGPRS: 56 -> 52 (-7.14 %) Spilled SGPRs: 0 -> 0 (0.00 %) Spilled VGPRs: 0 -> 0 (0.00 %) Private memory VGPRs: 0 -> 0 (0.00 %) Scratch size: 0 -> 0 (0.00 %) dwords per thread Code Size: 2168 -> 4560 (110.33 %) bytes LDS: 0 -> 0 (0.00 %) blocks Max Waves: 4 -> 4 (0.00 %) Wait states: 0 -> 0 (0.00 %) Reviewed-by: Thomas Helland <[email protected]> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=32211
* nir: reword code commentTimothy Arceri2018-12-131-2/+2
| | | | Reviewed-by: Thomas Helland <[email protected]>
* nir: in loop analysis track actual control flow typeTimothy Arceri2018-12-131-13/+21
| | | | | | | This will allow us to improve analysis to find more induction variables. Reviewed-by: Thomas Helland <[email protected]>
* nir: add if opt opt_if_loop_last_continue()Danylo Piliaiev2018-12-131-0/+95
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Removing the last continue can allow more loops to unroll. Also inserting code into the if branch can allow the various if opts to progress further. The insertion of some loops into the if branch also reduces VGPR use in some shaders. vkpipeline-db results (VEGA): Totals from affected shaders: SGPRS: 6552 -> 6576 (0.37 %) VGPRS: 6544 -> 6532 (-0.18 %) Spilled SGPRs: 0 -> 0 (0.00 %) Spilled VGPRs: 0 -> 0 (0.00 %) Private memory VGPRs: 0 -> 0 (0.00 %) Scratch size: 0 -> 0 (0.00 %) dwords per thread Code Size: 481952 -> 478032 (-0.81 %) bytes LDS: 13 -> 13 (0.00 %) blocks Max Waves: 241 -> 242 (0.41 %) Wait states: 0 -> 0 (0.00 %) Shader-db results radeonsi (VEGA): Totals from affected shaders: SGPRS: 168 -> 168 (0.00 %) VGPRS: 144 -> 140 (-2.78 %) Spilled SGPRs: 157 -> 157 (0.00 %) Spilled VGPRs: 0 -> 0 (0.00 %) Private memory VGPRs: 0 -> 0 (0.00 %) Scratch size: 0 -> 0 (0.00 %) dwords per thread Code Size: 8524 -> 8488 (-0.42 %) bytes LDS: 0 -> 0 (0.00 %) blocks Max Waves: 7 -> 7 (0.00 %) Wait states: 0 -> 0 (0.00 %) v2: (Timothy Arceri): - allow for continues in either branch - move any trailing loops inside the if as well as blocks. - leave nir_opt_trivial_continues() to actually remove the continue. Reviewed-by: Thomas Helland <[email protected]> Signed-off-by: Timothy Arceri <[email protected]> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=32211
* nir: rework force_unroll_array_access()Timothy Arceri2018-12-131-14/+35
| | | | | | | Here we rework force_unroll_array_access() so that we can reuse the induction variable detection in a following patch. Reviewed-by: Thomas Helland <[email protected]>
* nir: factor out some of the complex loop unroll code to a helperTimothy Arceri2018-12-131-51/+64
| | | | | Acked-by: Jason Ekstrand <[email protected]> Reviewed-by: Thomas Helland <[email protected]>
* nir: Document the function inlining processJason Ekstrand2018-12-121-0/+68
| | | | | | | | | This has thrown a few people off recently and it's good to have the process and all the rational for it documented somewhere. A comment at the top of nir_inline_functions seems as good a place as any. Acked-by: Karol Herbst <[email protected]> Reviewed-by: Matt Turner <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-27 17:30:42 +0000