aboutsummaryrefslogtreecommitdiffstats
path: root/src/glsl
Commit message (Collapse)AuthorAgeFilesLines
* nir: add an optimization to remove useless phi nodesConnor Abbott2015-02-033-0/+112
| | | | | | | | | | | | | | | | | | | | This removes phi nodes whose sources all point to the same thing. Shader-db results: total NIR instructions in shared programs: 2045293 -> 2041209 (-0.20%) NIR instructions in affected programs: 126564 -> 122480 (-3.23%) helped: 615 HURT: 0 total FS instructions in shared programs: 4321840 -> 4320392 (-0.03%) FS instructions in affected programs: 24622 -> 23174 (-5.88%) helped: 138 HURT: 0 Reviewed-by: Jason Ekstrand <[email protected]> Tested-by: Jason Ekstrand <[email protected]> Signed-off-by: Connor Abbott <[email protected]>
* nir/validate: Ensure that phi sources are SSA-onlyJason Ekstrand2015-02-031-10/+3
| | | | Reviewed-by: Connor Abbott <[email protected]>
* nir/validate: Validate that only float ALU outputs are saturatedJason Ekstrand2015-02-031-0/+8
| | | | Reviewed-by: Connor Abbott <[email protected]>
* nir/lower_source_mods: Don't lower saturate for non-float outputsJason Ekstrand2015-02-031-0/+4
| | | | Reviewed-by: Connor Abbott <[email protected]>
* nir: Add a pass to lower vector phi nodes to scalar phi nodesJason Ekstrand2015-02-033-0/+293
| | | | | | | | | | | | | | | | | | | | | | v2 Jason Ekstrand <[email protected]>: - Add better comments - Use nir_ssa_dest_init and nir_src_for_ssa more places - Fix some void * casts v3 Jason Ekstrand <[email protected]>: - Rework the way we determine whether or not to sccalarize a phi node to make the recursion non-bogus - Treat load_const instructions as scalarizable v4 Jason Ekstrand <[email protected]>: - Allow uniform and input loads to be scalarizable v5 Jason Ekstrand <[email protected]>: - Also consider loads of inputs (varying, uniform, or ubo) to be scalarizable. We were already doing this for load_var on uniforms and inputs. Reviewed-by: Kenneth Graunke <[email protected]>
* glsl/list: Note that exec_lists may not be realloc'd.Matt Turner2015-02-031-0/+4
| | | | Reviewed-by: Kenneth Graunke <[email protected]>
* glsl: Improve precision of mod(x,y)Iago Toral Quiroga2015-02-033-28/+41
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently, Mesa uses the lowering pass MOD_TO_FRACT to implement mod(x,y) as y * fract(x/y). This implementation has a down side though: it introduces precision errors due to the fract() operation. Even worse, since the result of fract() is multiplied by y, the larger y gets the larger the precision error we produce, so for large enough numbers the precision loss is significant. Some examples on i965: Operation Precision error ----------------------------------------------------- mod(-1.951171875, 1.9980468750) 0.0000000447 mod(121.57, 13.29) 0.0000023842 mod(3769.12, 321.99) 0.0000762939 mod(3769.12, 1321.99) 0.0001220703 mod(-987654.125, 123456.984375) 0.0160663128 mod( 987654.125, 123456.984375) 0.0312500000 This patch replaces the current lowering pass with a different one (MOD_TO_FLOOR) that follows the recommended implementation in the GLSL man pages: mod(x,y) = x - y * floor(x/y) This implementation eliminates the precision errors at the expense of an additional add instruction on some systems. On systems that can do negate with multiply-add in a single operation this new implementation would come at no additional cost. v2 (Ian Romanick) - Do not clone operands because when they are expressions we would be duplicating them and that can lead to suboptimal code. Fixes the following 16 dEQP tests: dEQP-GLES3.functional.shaders.builtin_functions.precision.mod.mediump_* dEQP-GLES3.functional.shaders.builtin_functions.precision.mod.highp_* Reviewed-by: Ian Romanick <[email protected]>
* glsl: can't have 'const' qualifier used with struct or interface block membersIago Toral Quiroga2015-02-031-0/+7
| | | | | | | | Fixes the following 2 dEQP tests: dEQP-GLES3.functional.shaders.declarations.invalid_declarations.uniform_block_const_vertex dEQP-GLES3.functional.shaders.declarations.invalid_declarations.uniform_block_const_fragment Reviewed-by: Ian Romanick <[email protected]>
* glsl: interface blocks must be declared at global scopeIago Toral Quiroga2015-02-031-0/+8
| | | | | | | | Fixes the following 2 dEQP tests: dEQP-GLES3.functional.shaders.declarations.invalid_declarations.uniform_block_in_main_vertex dEQP-GLES3.functional.shaders.declarations.invalid_declarations.uniform_block_in_main_fragment Reviewed-by: Ian Romanick <[email protected]>
* glsl: Pick ast_conditional branch regardless of op1/2 being constant.Kenneth Graunke2015-02-021-4/+2
| | | | | | | | | | | | | | | If the ?: operator's condition is a constant value, and both branches were pure expressions, we can just make the resulting value one or the other. Previously, we only did this if op[1] and op[2] were also constant values - but there's no actual reason for that restriction. No changes in shader-db, probably because we usually optimize this later anyway. But it does make us generate less stupid code up front. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* nir/opt_algebraic: Add some constant bcsel reductionsJason Ekstrand2015-01-291-2/+28
| | | | | | | | total instructions in shared programs: 5998190 -> 5997603 (-0.01%) instructions in affected programs: 54276 -> 53689 (-1.08%) helped: 293 Reviewed-by: Kenneth Graunke <[email protected]>
* nir/opt_algebraic: Add some boolean simplificationsJason Ekstrand2015-01-291-4/+5
| | | | | | | | total instructions in shared programs: 5998321 -> 5998287 (-0.00%) instructions in affected programs: 4520 -> 4486 (-0.75%) helped: 8 Reviewed-by: Kenneth Graunke <[email protected]>
* nir/algebraic: Support specifying variable as constant or by typeJason Ekstrand2015-01-292-6/+26
| | | | Reviewed-by: Kenneth Graunke <[email protected]>
* nir/algebraic: Fail to compile of a variable is used in a replace but not ↵Jason Ekstrand2015-01-291-0/+7
| | | | | | the search Reviewed-by: Kenneth Graunke <[email protected]>
* nir/search: Allow for matching variables based on typesJason Ekstrand2015-01-292-0/+23
| | | | | | | | | | | This allows you to match on an unknown value but only if it is of a given type. 90% of the uses of this are for matching only booleans, but adding the generality of arbitrary types is no more complex. nir_algebraic.py doesn't handle this yet but that's ok because the C language will ensure that the default type on all variables is void. Reviewed-by: Kenneth Graunke <[email protected]>
* nir/search: Add support for matching unknown constantsJason Ekstrand2015-01-292-0/+13
| | | | | | | | | | | | | There are some algebraic transformations that we want to do but only if certain things are constants. For instance, we may want to replace a * (b + c) with (a * b) + (a * c) as long as a and either b or c is constant. While this generates more instructions, some of it will get constant folded. nir_algebraic.py doesn't handle this yet, but that's ok because the C language will make sure that false is the default for now. Reviewed-by: Kenneth Graunke <[email protected]>
* nir: Add an invalid typeJason Ekstrand2015-01-291-0/+1
| | | | | | This allows us to indicate a concept of an invalid type. Reviewed-by: Kenneth Graunke <[email protected]>
* nir: Add variants of some of the comparison simplifications.Eric Anholt2015-01-291-0/+4
| | | | | | | | | | | We end up with these from TGSI-to-NIR because the pass generating the comparisons doesn't know if the arg is actually a bool input or not. vc4 results: total instructions in shared programs: 41801 -> 41508 (-0.70%) instructions in affected programs: 4253 -> 3960 (-6.89%) Reviewed-by: Matt Turner <[email protected]>
* nir: Don't try to to-SSA ALU instructions that are already SSA.Eric Anholt2015-01-291-0/+3
| | | | Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Fix a bit of broken indentation.Eric Anholt2015-01-291-1/+1
| | | | | Reviewed-by: Matt Turner <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Add a couple of helpers for glsl types.Eric Anholt2015-01-292-1/+16
| | | | | | | | | | This will be used by tgsi_to_nir, which needs to get vec4 types for declaring shader input/output variables. v2: Add a missing space. Reviewed-by: Matt Turner <[email protected]> (v2) Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Make vec-to-movs handle src/dest aliasing.Eric Anholt2015-01-281-10/+72
| | | | | | | | | | | | | | | It now emits vector MOVs instead of a series of individual MOVs, which should be useful to any vector backends. This pushes the problem of src/dest aliasing of channels on a scalar chip to the backend, but if there are any vector operations in your shader then you needed to be handling this already. Fixes fs-swap-problem with my scalarizing patches. v2: Rename to insert_mov(), and add a comment about what it does. v3: Rewrite the comment. Reviewed-by: Connor Abbott <[email protected]> (v3)
* nir/opcodes: Use a return type of tfloat for ldexpJason Ekstrand2015-01-281-1/+1
| | | | | Reviewed-by: Connor Abbott <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* Revert "nir/opcodes: Use fpclassify() instead of isnormal() for ldexp"Jason Ekstrand2015-01-281-1/+1
| | | | | | | | | | | This reverts commit d7d340fb2f68c46bd5a0008ecf53c6693e29c916. We have an isnormal() implementation available, the only problem was that we had the wrong return type (fixed in a later patch). Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88806 Acked-by: Matt Turner <[email protected]>
* nir/opcodes: Use fpclassify() instead of isnormal() for ldexpJason Ekstrand2015-01-281-1/+1
| | | | | Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88806 Reviewed-by: Ian Romanick <[email protected]>
* nir: fix a bug with constant folding non-per-component instructionsConnor Abbott2015-01-261-1/+2
| | | | | | | | | | | | | Before, we were only copying the first N channels, where N is the size of the SSA destination, which is fine for per-component instructions, but non-per-component instructions like fdot3 can have more source components than destination components. Fix this using the helper function introduced in the last patch. v2: use new helper name Reviewed-by: Jason Ekstrand <[email protected]> Signed-off-by: Connor Abbott <[email protected]>
* nir: add a helper function for getting the number of source componentsConnor Abbott2015-01-261-0/+15
| | | | | | | | | | | | | | Unlike with non-SSA ALU instructions, where if they're per-component you have to look at the writemask to know which source channels are being used, SSA ALU instructions always have all the possible channels enabled so we can just look at the number of components in the SSA definition for per-component instructions to say how many source components are being used. v2: use new name nir_ssa_alu_instr_src_components() Reviewed-by: Jason Ekstrand <[email protected]> Signed-off-by: Connor Abbott <[email protected]>
* nir/opcodes: Don't go through doubles when constant-folding iabsJason Ekstrand2015-01-261-1/+1
| | | | | | | | | Previously, we called the abs() function in math.h. However, this involves unnecessarily going through double. This commit changes it to use integers directly with a ternary. Reviewed-by: Connor Abbott <[email protected]> Reviewed-by: Eric Anholt <[email protected]>
* nir/opcodes: Simplify and fix the unpack_half_*_split_* constant expressionsJason Ekstrand2015-01-261-6/+4
| | | | | | | | | | Previously, these functions were explicitly writing to dst.x and dst.y. However they both return only one component so writing to dst.y is invalid. Also, since they only return one component, we don't need the explicit assignment in the expression and can simplify it use an implicit assignment. Reviewed-by: Connor Abbott <[email protected]>
* nir: Use pointers for nir_src_copy and nir_dest_copyJason Ekstrand2015-01-2610-53/+47
| | | | | | | | This avoids the overhead of copying structures and better matches the newly added nir_alu_src_copy and nir_alu_dest_copy. Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Connor Abbott <[email protected]>
* nir/constant_folding: use the new constant folding infrastructureConnor Abbott2015-01-241-158/+21
| | | | | Signed-off-by: Connor Abbott <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir: add new constant folding infrastructureJason Ekstrand2015-01-246-184/+787
| | | | | | | | | | | | | | | | | | | | | | | | | Add a required field to the Opcode class, const_expr, that contains an expression or statement that computes the result of the opcode given known constant inputs. Then take those const_expr's and expand them into a function that takes an opcode and an array of constant inputs and spits out the constant result. This means that when adding opcodes, there's one less place to update, and almost all the opcodes are self-documenting since the information on how to compute the result is right next to the definition. The helper functions in nir_constant_expressions.c were taken from ir_constant_expressions.cpp. v3 Jason Ekstrand <[email protected]> - Use mako to generate one function per opcode instead of doing piles of string splicing v4 Jason Ekstrand <[email protected]> - More comments and better indentation in the mako - Add a description of the constant expression language in nir_opcodes.py - Added nir_constant_expressions.py to EXTRA_DIST in Makefile.am Signed-off-by: Jason Ekstrand <[email protected]> Reviewed-by: Connor Abbott <[email protected]>
* nir: use Python to autogenerate opcode informationConnor Abbott2015-01-248-401/+478
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Before, we used a system where a file, nir_opcodes.h, defined some macros that were included to generate the enum values and the nir_op_infos structure. This worked pretty well, but for development the error messages were never very useful, Python tools couldn't understand the opcode list, and it was difficult to use nir_opcodes.h to do other things like autogenerate a builder API. Now, we store opcode information in nir_opcodes.py, and we have nir_opcodes_c.py to generate the old nir_opcodes.c and nir_opcodes_h.py to generate nir_opcodes.h, which contains all the enum names and gets included into nir.h like before. In addition to solving the above problems, using Python and Mako to generate everything means that it's much easier to add keep information centralized as we add new things like constant propagation that require per-opcode information. v2: - make Opcode derive from object (Dylan) - don't use assert like it's a function (Dylan) - style fixes for fnoise, use xrange (Dylan) - use iterkeys() in nir_opcodes_h.py (Dylan) - use pydoc-style comments (Jason) - don't make fmin/fmax commutative and associative yet (Jason) Signed-off-by: Connor Abbott <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> v3 Jason Ekstrand <[email protected]> - Alphabetize source file lists - Generate nir_opcodes.h in the builddir instead of the source dir - Include $(builddir)/src/glsl/nir in the i965 build - Rework nir_opcodes.h generation so it generates a complete header file instead of one that has to be embedded inside an enum declaration
* glsl: Add a foreach_in_list_reverse_safe macro.Matt Turner2015-01-231-0/+6
| | | | Reviewed-by: Kenneth Graunke <[email protected]>
* nir: Expose nir_print_instr() for debug printsEric Anholt2015-01-232-2/+8
| | | | | | | | | It's nice to have this present in your default cases so you can see what instruction is triggering an abort. v2: Just pass a NULL state, now that it won't crash when you do. Reviewed-by: Jason Ekstrand <[email protected]>
* nir: When asked to print with a NULL state, just use bare variable names.Eric Anholt2015-01-231-6/+16
| | | | Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Add nir_lower_alu_to_scalar.Eric Anholt2015-01-233-0/+188
| | | | | | | | | | | | | | | | This is the equivalent of brw_fs_channel_expressions.cpp, which I wanted for vc4. v2: Use the nir_src_for_ssa() helper, and another instance of nir_alu_src_copy(). v3: Drop the non-SSA support. All intended callers will have SSA-only ALU ops. v4: Use insert_before, drop stale bcsel/fcsel comment, drop now-unused unsupported() function, drop lower_context struct. v5: Completely rename the pass to nir_lower_alu_to_scalar(), add an assert about weird input_sizes[]. Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Make some helpers for copying ALU src/dests.Eric Anholt2015-01-234-9/+25
| | | | | | | | | There aren't many users yet, but I wanted to do this from my scalarizing pass. v2: Constify the src arguments. Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Add algebraic optimizations for division and reciprocal.Kenneth Graunke2015-01-231-0/+5
| | | | | | | | | | | | | | | | | These also exist in opt_algebraic.cpp. total NIR instructions in shared programs: 2011430 -> 2011211 (-0.01%) NIR instructions in affected programs: 42221 -> 42002 (-0.52%) helped: 198 total i965 instructions in shared programs: 6020553 -> 6020116 (-0.01%) i965 instructions in affected programs: 84322 -> 83885 (-0.52%) helped: 394 HURT: 1 (by 1 instruction) Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* nir: Add algebraic optimizations for exponential/logarithmic functions.Kenneth Graunke2015-01-231-0/+10
| | | | | | | | | | | | | | | | | | Most of these exist in the GLSL IR algebraic pass already. However, SSA allows us to find more instances of the patterns. total NIR instructions in shared programs: 2015593 -> 2011430 (-0.21%) NIR instructions in affected programs: 124189 -> 120026 (-3.35%) helped: 604 total i965 instructions in shared programs: 6025505 -> 6018717 (-0.11%) i965 instructions in affected programs: 261295 -> 254507 (-2.60%) helped: 1295 HURT: 3 Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* nir: Add algebraic optimizations for simplifying comparisons.Kenneth Graunke2015-01-231-0/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | The first batch removes bonus fnot/inot operations, possibly allowing other optimizations to better recognize patterns. The next batch replaces a fadd and constant 0.0 with an fneg - negation is usually free on GPUs, while addition is not. total NIR instructions in shared programs: 2020814 -> 2015593 (-0.26%) NIR instructions in affected programs: 411143 -> 405922 (-1.27%) helped: 2233 HURT: 214 A few shaders are hurt by a few instructions due to moving neg such that it has a constant operand, which is then folded, resulting in two distinct load_consts for x and -x. We can always clean that up later. total i965 instructions in shared programs: 6035392 -> 6025505 (-0.16%) i965 instructions in affected programs: 784980 -> 775093 (-1.26%) helped: 4508 HURT: 2 Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* nir: Add algebraic optimizations for pointless shifts.Kenneth Graunke2015-01-231-0/+7
| | | | | | | | | | | | | | | | | The GLSL IR optimization pass contained these; we may as well include them too. v2: Fix a >> 0 and a << 0 optimizations (caught by Matt). No change in the number of NIR instructions on a shader-db run. total i965 instructions in shared programs: 6035397 -> 6035392 (-0.00%) i965 instructions in affected programs: 542 -> 537 (-0.92%) helped: 2 (in glamor) Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* nir: Add a bunch of algebraic optimizations on logic/bit operations.Kenneth Graunke2015-01-231-0/+13
| | | | | | | | | | | | | | | | | | | | Matt and I noticed a bunch of "val <- ior a a" operations in a shader, so we decided to add an algebraic optimization for that. While there, I decided to add a bunch more of them. v2: Delete bogus fand/for optimizations (caught by Jason). total NIR instructions in shared programs: 2023511 -> 2020814 (-0.13%) NIR instructions in affected programs: 149634 -> 146937 (-1.80%) helped: 1032 total i965 instructions in shared programs: 6035392 -> 6035397 (0.00%) i965 instructions in affected programs: 537 -> 542 (0.93%) HURT: 2 Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* nir: Implement CSE on intrinsics that can be eliminated and reordered.Kenneth Graunke2015-01-231-2/+38
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Matt and I noticed that one of the shaders hurt by INTEL_USE_NIR=1 had load_input and load_uniform intrinsics repeated several times, with the same parameters, but each one generating a distinct SSA value. This made ALU operations on those values appear distinct as well. Generating distinct SSA values is silly - these are read only variables. CSE'ing them makes everything use a single SSA value, which then allows other operations to be CSE'd away as well. Generalizing a bit, it seems like we should be able to safely CSE any intrinsics that can be eliminated and reordered. I didn't implement support for variables for the time being. v2: Assert that info->num_variables == 0 (requested by Jason). total NIR instructions in shared programs: 2435936 -> 2023511 (-16.93%) NIR instructions in affected programs: 2413496 -> 2001071 (-17.09%) helped: 16872 total i965 instructions in shared programs: 6028987 -> 6008427 (-0.34%) i965 instructions in affected programs: 640654 -> 620094 (-3.21%) helped: 2071 HURT: 585 GAINED: 14 LOST: 25 Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
* nir: Pull nir_instr_can_cse()'s SSA checks out of the switch.Kenneth Graunke2015-01-231-2/+6
| | | | | | | | | | | This should not be a change in behavior, as all current cases that potentially answer "yes" require SSA. The next patch will introduce another case that requires SSA. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* glsl: Build a libglsl_util library.Matt Turner2015-01-231-16/+7
| | | | | Rather than sourcing files with ../dir/file.c which leads to distclean wiping out ../dir's .deps directory.
* glsl: Build with subdir-objects.Matt Turner2015-01-233-190/+188
| | | | | | Apparently $(top_srcdir) is not expanded in a source list when using subdir-objects, so remove that. It's not clear to me why we were going to such lengths to prefix each source file anyway.
* nir: Add headers to distribution.Matt Turner2015-01-231-0/+2
|
* nir: Add nir_{opt_,}algebraic.py to distribution.Matt Turner2015-01-231-0/+2
|
* nir: add generated file to .gitignoreConnor Abbott2015-01-231-0/+1
| | | | | Signed-off-by: Connor Abbott <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-05 02:49:20 +0000