Steven Perron [Wed, 14 Feb 2018 02:08:43 +0000 (21:08 -0500)]
Simplify OpPhi instructions referencing unreachable continues
In dead branch elimination, we already recognize unreachable continue
blocks, and update OpPhi instruction accordingly. This change adds an
extra check: if the head block has exactly 1 other incoming edge, then
replace the OpPhi with the value from that edge.
Fixes #1314.
Steven Perron [Thu, 15 Feb 2018 15:41:01 +0000 (10:41 -0500)]
Get CCP to use the constant floating point rules.
Fixes #1311
Lei Zhang [Fri, 16 Feb 2018 00:40:47 +0000 (19:40 -0500)]
SubgroupBallotKHR can enable SubgroupSize & SubgroupLocalInvocationId
Nerijus Baliūnas [Wed, 14 Feb 2018 19:06:53 +0000 (21:06 +0200)]
fix typo
David Neto [Wed, 14 Feb 2018 23:27:22 +0000 (18:27 -0500)]
Invoke cmake via CMAKE_COMMAND variable
Need to do this in case cmake is not on the path.
This should fix the Android NDK build, as in when building the NDK
itself.
Lei Zhang [Thu, 15 Feb 2018 15:19:15 +0000 (10:19 -0500)]
Avoid using static unordered_map (#1304)
unordered_map is not POD. Using it as static may cause problems
when operator new() and operator delete() is customized.
Also changed some function signatures to use const char* instead
of std::string, which will give caller the flexibility to avoid
creating a std::string.
Arseny Kapoulkine [Thu, 15 Feb 2018 04:45:36 +0000 (20:45 -0800)]
Add folding of redundant OpSelect insns
We can fold OpSelect into one of the operands in two cases:
- condition is constant
- both results are the same
Even if the original shader doesn't have either of these, if-conversion
pass sometimes ends up generating instructions like
%7127 = OpSelect %int %3220 %7058 %7058
And this optimization cleans them up.
Steven Perron [Tue, 13 Feb 2018 04:28:38 +0000 (23:28 -0500)]
Add id to name map
Adding a map from an id to it set of OpName and OpMemberName
instructions. This will be used in KillNameAndDecorates to kill the
names for the ids that are being removed.
In my test, the compile time for 50 shaders went from 1m57s to 55s.
This was on linux using the release build.
Fixes #1290.
Steven Perron [Fri, 9 Feb 2018 18:37:26 +0000 (13:37 -0500)]
Fold binary floating point operators.
Adds the floating rules for FAdd, FDiv, FMul, and FSub.
Contributes to #1164.
Stephen McGroarty [Wed, 14 Feb 2018 17:03:12 +0000 (17:03 +0000)]
Initial support for loop unrolling.
This patch adds initial support for loop unrolling in the form of a
series of utility classes which perform the unrolling. The pass can
be run with the command spirv-opt --loop-unroll. This will unroll
loops within the module which have the unroll hint set. The unroller
imposes a number of requirements on the loops it can unroll. These are
documented in the comments for the LoopUtils::CanPerformUnroll method in
loop_utils.h. Some of the restrictions will be lifted in future patches.
Alan Baker [Wed, 14 Feb 2018 14:39:11 +0000 (09:39 -0500)]
Fixes #1295. Mark undef values as varying in ccp.
* Undef now marked as varying in ccp
* this prevents incorrect meet operations since phis were always not
interesting
* added a test to catch the bug
Diego Novillo [Mon, 12 Feb 2018 18:21:48 +0000 (13:21 -0500)]
Cleanup. Use proper #include guard. NFC.
Steven Perron [Sat, 10 Feb 2018 01:33:06 +0000 (20:33 -0500)]
Avoid using the def-use manager during inlining.
There seems to only be a single location where the def-use manager is
used. It is to get information about a type. We can do that with the
type manager instead.
Fixes #1285
Arseny Kapoulkine [Sat, 10 Feb 2018 01:57:29 +0000 (17:57 -0800)]
Fix spirv.h include to rely on include paths
This is important when SPIRV-Headers are not checked out to external/
folder and mirrors other places in the code where spirv.h is included.
Steven Perron [Thu, 8 Feb 2018 15:59:03 +0000 (10:59 -0500)]
Add folding of OpCompositeExtract and OpConstantComposite constant instructions.
Create files for constant folding rules.
Add the rules for OpConstantComposite and OpCompositeExtract.
David Neto [Thu, 8 Feb 2018 17:05:40 +0000 (12:05 -0500)]
Fix generation of Vim syntax file
Józef Kucia [Fri, 9 Feb 2018 11:30:08 +0000 (12:30 +0100)]
Do not hardcode libdir and includedir in pkg config files
Steven Perron [Thu, 8 Feb 2018 16:03:48 +0000 (11:03 -0500)]
Add header files missing from CMakeLists.txt
Alexander Johnston [Mon, 29 Jan 2018 10:39:55 +0000 (10:39 +0000)]
Loop invariant code motion initial implementation
GregF [Wed, 7 Feb 2018 02:54:58 +0000 (19:54 -0700)]
SROA: Do replacement on structs with no partial references.
Steven Perron [Fri, 2 Feb 2018 16:55:05 +0000 (11:55 -0500)]
Make use of the instruction folder.
Implementation of the simplification pass.
- Create pass that calls the instruction folder on each instruction and
propagate instructions that fold to a copy. This will do copy
propagation as well.
- Did not use the propagator engine because I want to modify the instruction
as we go along.
- Change folding to not allocate new instructions, but make changes in
place. This change had a big impact on compile time.
- Add simplification pass to the legalization passes in place of
insert-extract elimination.
- Added test cases for new folding rules.
- Added tests for the simplification pass
- Added a method to the CFG to apply a function to the basic blocks in
reverse post order.
Contributes to #1164.
Andrey Tuganov [Tue, 6 Feb 2018 20:25:03 +0000 (15:25 -0500)]
Disable check which fails Vulkan CTS
Andrey Tuganov [Tue, 6 Feb 2018 17:10:11 +0000 (12:10 -0500)]
Add Vulkan-specific validation rules for atomics
Added atomic instructions validation rules from
https://www.khronos.org/registry/vulkan/specs/1.0/html/vkspec.html#spirvenv-module-validation
Józef Kucia [Fri, 2 Feb 2018 22:37:14 +0000 (23:37 +0100)]
Build SPIRV-Tools as shared library
Add pkg-config file for shared libraries
Properly build SPIRV-Tools DLL
Test C interface with shared library
Set PATH to shared library file for c_interface_shared test
Otherwise, the test won't find SPIRV-Tools-shared.dll.
Do not use private functions when testing with shared library
Make all symbols hidden by default for shared library target
David Neto [Tue, 6 Feb 2018 16:13:39 +0000 (08:13 -0800)]
Enable Visual Studio 2013 again
Disable use of Effcee and RE2 with MSVC compilers older than Visual
Studio 2015 since RE2 doesn't support them.
David Neto [Tue, 6 Feb 2018 17:06:26 +0000 (12:06 -0500)]
Fix test inclusion when Effcee is absent
David Neto [Tue, 6 Feb 2018 16:47:44 +0000 (11:47 -0500)]
Update CHANGES
Alan Baker [Mon, 5 Feb 2018 14:47:05 +0000 (09:47 -0500)]
Registering a type now rebuilds it out of memory owned by the manager.
* Added TypeManager::RebuildType
* rebuilds the type and its constituent types in terms of memory owned
by the manager.
* Used by TypeManager::RegisterType to properly allocate memory
* Adding an unit test to expose the issue
* Added some tests to provide coverage of RebuildType
* Added an accessor to the target pointer for a forward pointer
GregF [Fri, 2 Feb 2018 23:57:39 +0000 (16:57 -0700)]
ADCE: Fix combinator initialization
The combinator initialization was only looking at the capabilities
in the shader and not the inferred capabilities. Geometry and tessellation
shaders were not setting the Shader capability which is inferred. So the
combinator set was not initialized correctly causing problems for ADCE.
David Neto [Mon, 5 Feb 2018 15:56:45 +0000 (07:56 -0800)]
VS2013: LoopDescriptor LoopContainerType can't contain unique_ptr
The loop descriptor must explicitly manage the storage for contained
Loop objects.
Fixes #1262
Andrey Tuganov [Wed, 31 Jan 2018 21:29:54 +0000 (16:29 -0500)]
Add barrier instructions validation pass
David Neto [Fri, 2 Feb 2018 15:28:52 +0000 (10:28 -0500)]
Avoid vector copies in range-for loops in opt/types.cpp
Also be more explicit about iterated types in other range-for loops.
David Neto [Fri, 2 Feb 2018 22:17:42 +0000 (14:17 -0800)]
Disambiguate between const and nonconst ForEachSuccessorLabel
This helps VisualStudio 2013 compile the code.
Contributes to #1262
Steven Perron [Wed, 24 Jan 2018 18:26:33 +0000 (13:26 -0500)]
Add general folding infrastructure.
Create the folding engine that will
1) attempt to fold an instruction.
2) iterates on the folding so small folding rules can be easily combined.
3) insert new instructions when needed.
I've added the minimum number of rules needed to test the features above.
David Neto [Fri, 2 Feb 2018 16:27:07 +0000 (11:27 -0500)]
Start v2018.1-dev
David Neto [Fri, 2 Feb 2018 16:25:58 +0000 (11:25 -0500)]
Finalize v2018.0
David Neto [Fri, 2 Feb 2018 16:23:33 +0000 (11:23 -0500)]
Update CHANGES
Alan Baker [Wed, 31 Jan 2018 15:40:33 +0000 (10:40 -0500)]
Reordering performance passes ordering to produce better opts
* Moved initial insert/extract passes later to cover more opportunities
* Added an extra set of passes to clean up opportunities exposed later
in the pipeline
Victor Lomuller [Fri, 26 Jan 2018 12:07:10 +0000 (12:07 +0000)]
Add LoopUtils class to gather some loop transformation support.
This patch adds LoopUtils class to handle some loop related transformations. For now it has 2 transformations that simplifies other transformations such as loop unroll or unswitch:
- Dedicate exit blocks: this ensure that all exit basic block
(out-of-loop basic blocks that have a predecessor in the loop)
have all their predecessors in the loop;
- Loop Closed SSA (LCSSA): this ensure that all definitions in a loop are used inside the loop
or in a phi instruction in an exit basic block.
It also adds the following capabilities:
- Loop::IsLCSSA to test if the loop is in a LCSSA form
- Loop::GetOrCreatePreHeaderBlock that can build a loop preheader if required;
- New methods to allow on the fly updates of the loop descriptors.
- New methods to allow on the fly updates of the CFG analysis.
- Instruction::SetOperand to allow expression of the index relative to Instruction::NumOperands (to be compatible with the index returned by DefUseManager::ForEachUse)
Steven Perron [Tue, 30 Jan 2018 16:24:03 +0000 (11:24 -0500)]
Add pass to reaplce invalid opcodes
Creates a pass that will remove instructions that are invalid for the
current shader stage. For the instruction to be considered for replacement
1) The opcode must be valid for a shader modules.
2) The opcode must be invalid for the current shader stage.
3) All entry points to the module must be for the same shader stage.
4) The function containing the instruction must be reachable from an entry point.
Fixes #1247.
Andrey Tuganov [Tue, 9 Jan 2018 15:58:22 +0000 (10:58 -0500)]
Added OpenCL ExtInst validation rules
Jeremy Hayes [Thu, 25 Jan 2018 23:32:01 +0000 (16:32 -0700)]
Add adjacency validation pass
Validate OpPhi predecessors.
Validate OpLoopMerge successors.
Validate OpSelectionMerge successors.
Fix collateral damage to existing tests.
Remove ValidateIdWithMessage.OpSampledImageUsedInOpPhiBad.
Andrey Tuganov [Wed, 31 Jan 2018 21:59:55 +0000 (16:59 -0500)]
Fixed harmless uninit var warning
David Neto [Tue, 23 Jan 2018 09:50:54 +0000 (17:50 +0800)]
Use SPIR-V headers from "unified1" directory
Alan Baker [Tue, 30 Jan 2018 21:43:55 +0000 (16:43 -0500)]
Remove constexpr from Analysis operators
* Had to remove templating from InstructionBuilder as a result
* now preserved analyses are specified as a constructor argument
* updated tests and uses
* changed static_assert to a runtime assert
* this should probably get further changes in the future
GregF [Wed, 31 Jan 2018 01:13:30 +0000 (18:13 -0700)]
Opt: Add ScalarReplacement to RegisterSizePasses
Andrey Tuganov [Thu, 25 Jan 2018 18:39:43 +0000 (13:39 -0500)]
Add memory semantics checks to validate atomics
David Neto [Tue, 30 Jan 2018 22:47:00 +0000 (17:47 -0500)]
Update CHANGES
Alan Baker [Tue, 30 Jan 2018 20:00:27 +0000 (15:00 -0500)]
Prevent unnecessary changes to the IR in dead branch elim
* When handling unreachable merges and continues, do not optimize to the
same IR
* pass did not check whether the unreachable blocks were in the
optimized form before transforming them
* added a test to catch this issue
Andrey Tuganov [Fri, 26 Jan 2018 16:29:15 +0000 (11:29 -0500)]
Improved error message in val capabilities
Alan Baker [Thu, 25 Jan 2018 22:40:06 +0000 (14:40 -0800)]
Enhancements to block merging
* Should handle all possibilities
* Stricter checks for what is disallowed:
* header and header
* merge and merge
* Allow header and merge blocks to be merged
* Erases the structured control declaration if merging header and
merge blocks together.
Alan Baker [Tue, 30 Jan 2018 15:15:43 +0000 (10:15 -0500)]
Fix dereference of possibly nullptr
* If the dead branch elim is performed on a module without structured
control flow, the OpSelectionMerge may not be present
* Add a check for pointer validity before dereferencing
* Added a test to catch the bug
GregF [Sat, 27 Jan 2018 00:05:33 +0000 (17:05 -0700)]
InsertExtractElim: Split out DeadInsertElim as separate pass
Alan Baker [Thu, 25 Jan 2018 18:45:21 +0000 (10:45 -0800)]
Fixes in CCP for #1228
* Forces traversal of phis if the def has changed to varying
* Mark a phi as varying if all incoming values are varying
* added a test to catch the bug
Victor Lomuller [Thu, 25 Jan 2018 10:33:06 +0000 (10:33 +0000)]
Add LoopDescriptor as an IRContext analysis.
Move some function definitions from header to source to avoid circular definition.
Greg Fischer [Thu, 11 Jan 2018 23:23:58 +0000 (16:23 -0700)]
DeadInsertElim: Detect and DCE dead Inserts
This adds Dead Insert Elimination to the end of the
--eliminate-insert-extract pass. See the new tests for examples of code
that will benefit.
Essentially, this removes OpCompositeInsert instructions which are not
used, either because there is no instruction which uses the value at the
index it is inserted, or because a subsequent insert intercepts any such
use.
This code has been seen to remove significant amounts of dead code from
real-life HLSL shaders being ported to Vulkan. In fact, it is needed to
remove dead texture samples which cause Vulkan validation layer errors
(unbound textures and samplers) if not removed . Such DCE is thus
required for fxc equivalence and legalization.
This analysis operates across "chains" of Inserts which can also contain
Phi instructions.
Alan Baker [Tue, 16 Jan 2018 16:15:06 +0000 (11:15 -0500)]
Initial implementation of if conversion
* Handles simple cases only
* Identifies phis in blocks with two predecessors and attempts to
convert the phi to an select
* does not perform code motion currently so the converted values must
dominate the join point (e.g. can't be defined in the branches)
* limited for now to two predecessors, but can be extended to handle
more cases
* Adding if conversion to -O and -Os
Andrey Tuganov [Wed, 24 Jan 2018 17:48:55 +0000 (12:48 -0500)]
Validator: restricted some atomic ops for shaders
Ban floating point case for OpAtomicLoad, OpAtomicExchange,
OpAtomicCompareExchange. In graphics (Shader) environments, these
instructions only operate on scalar integers. Ban the floating point
case. OpenCL supports atomic_float.
Andrey Tuganov [Tue, 23 Jan 2018 20:02:27 +0000 (15:02 -0500)]
Added Vulkan-specifc checks to image validation
Implemented Vulkan-specific rules:
- OpTypeImage must declare a scalar 32-bit float or 32-bit integer type
for the “Sampled Type”.
- OpSampledImage must only consume an “Image” operand whose type has its
“Sampled” operand set to 1.
Steven Perron [Mon, 22 Jan 2018 19:48:43 +0000 (14:48 -0500)]
Use id_map in Fold*ToConstant
The folding routines are suppose to use the id_map provided to map the
ids in the instruction. The ones I just added are missing it.
Steven Perron [Tue, 9 Jan 2018 17:45:46 +0000 (12:45 -0500)]
Add generic folding function and use in CCP
The current folding routines have a very cumbersome interface, make them
harder to use, and not a obvious how to extend.
This change is to create a new interface for the folding routines, and
show how it can be used by calling it from CCP.
This does not make a significant change to the behaviour of CCP. In
general it should produce the same code as before; however it is
possible that an instruction that takes 32-bit integers as inputs and
the result is not a 32-bit integer or bool will not be folded as before.
It seems like andriod has a problem with INT32_MAX and the like. I'll
explicitly define those if the are not already defined.
Alan Baker [Thu, 18 Jan 2018 20:44:00 +0000 (15:44 -0500)]
Fixes infinite loop in ADCE
* Addresses how breaks are indentified to prevent infinite loops when
back to back loop share a merge and header
* Added test to catch the bug
Victor Lomuller [Thu, 18 Jan 2018 17:14:58 +0000 (17:14 +0000)]
Introduce an instruction builder helper class.
The class factorize the instruction building process.
Def-use manager analysis can be updated on the fly to maintain coherency.
To be updated to take into account more analysis.
Alan Baker [Thu, 18 Jan 2018 19:15:54 +0000 (14:15 -0500)]
Simplifying code for adding instructions to worklist
* AddToWorklist can now be called unconditionally
* It will only add instructions that have not already been marked as
live
* Fixes a case where a merge was not added to the worklist because the
branch was already marked as live
* Added two similar tests that fail without the fix
Steven Perron [Wed, 17 Jan 2018 19:57:37 +0000 (14:57 -0500)]
Create a pass to work around a driver bug related to OpUnreachable.
We have come across a driver bug where and OpUnreachable inside a loop
is causing the shader to go into an infinite loop. This commit will try
to avoid this bug by turning OpUnreachable instructions that are
contained in a loop into branches to the loop merge block.
This is not added to "-O" and "-Os" because it should only be used if
the driver being targeted has this problem.
Fixes #1209.
Alan Baker [Thu, 18 Jan 2018 16:02:57 +0000 (11:02 -0500)]
Adding testcase for #1210
Victor Lomuller [Thu, 18 Jan 2018 08:51:12 +0000 (08:51 +0000)]
CFG: force the creation of a predecessor entry for all basic block.
This ensure that all basic blocks in a function have a valid entry the CFG object.
The entry block has no predecessors but remains a valid basic block
for which we might want to query the number of predecessors.
Some unreachable basic blocks may not have predecessors as well.
David Neto [Wed, 17 Jan 2018 15:57:20 +0000 (10:57 -0500)]
spirv-dis: Add --color option to force color disassembly
Fixes https://github.com/KhronosGroup/SPIRV-Tools/issues/1170
Alan Baker [Wed, 17 Jan 2018 19:09:24 +0000 (14:09 -0500)]
Fixing missing early exit from break identification
Alan Baker [Mon, 15 Jan 2018 18:25:45 +0000 (13:25 -0500)]
Adding support for switch removal in ADCE
* Updated code to handle switches
* Enabled disabled test and added a couple new ones
Alan Baker [Tue, 16 Jan 2018 21:27:13 +0000 (16:27 -0500)]
Capturing value table by reference in local redundancy elim
David Neto [Wed, 17 Jan 2018 03:52:39 +0000 (22:52 -0500)]
Update CHANGES
Alan Baker [Tue, 16 Jan 2018 19:47:35 +0000 (14:47 -0500)]
Fixes missing increment in common uniform elim
* Addresses #1203
* Increments inIdx in IsConstantIndexAccessChain
* added test to catch the bug
Steven Perron [Sat, 13 Jan 2018 05:28:40 +0000 (00:28 -0500)]
Skip SpecConstants in CCP.
At the moment specialization constants look like constants to ccp. This
causes a problem because they are handled differently by the constant
manager.
I choose to simply skip over them, and not try to add them to the value
table. We can do specialization before ccp if we want to be able to
propagate these values.
Fixes #1199.
David Neto [Fri, 12 Jan 2018 23:52:42 +0000 (18:52 -0500)]
Update CHANGES
Greg Fischer [Fri, 12 Jan 2018 22:52:45 +0000 (15:52 -0700)]
Add MatrixConstant
Steven Perron [Fri, 12 Jan 2018 20:15:57 +0000 (15:15 -0500)]
Remove redundant passes from legalization passes
With work that Alan has done, some passes have become redundant. ADCE
now removed unused variables. Dead branch elimination removes
unreachable blocks. This means we can remove CFG Cleanup and dead
variable elimination.
Alan Baker [Fri, 12 Jan 2018 20:05:53 +0000 (15:05 -0500)]
Adding early exit versions of several ForEach* methods
* Looked through code for instances where code would benefit from early
exit
* Added a corresponding WhileEach* method and updated the code
Steven Perron [Fri, 12 Jan 2018 16:23:57 +0000 (11:23 -0500)]
Move initialization of the const mgr to the constructor.
The current code expects the users of the constant manager to initialize
it with all of the constants in the module. The problem is that you do
not want to redo the work multiple times. So I decided to move that
code to the constructor of the constant manager. This way it will
always be initialized on first use.
I also removed an assert that expects all constant instructions to be
successfully mapped. This is because not all OpConstant* instruction
can map to a constant, and neither do the OpSpecConstant* instructions.
The real problem is that an OpConstantComposite can contain a member
that is OpUndef. I tried to treat OpUndef like OpConstantNull, but this
failed because an OpSpecConstantComposite with an OpUndef cannot be
changed to an OpConstantComposite. Since I feel this case will not be
common, I decided to not complicate the code.
Fixes #1193.
David Neto [Fri, 12 Jan 2018 18:33:43 +0000 (13:33 -0500)]
Start v2018.0-dev
David Neto [Fri, 12 Jan 2018 18:32:04 +0000 (13:32 -0500)]
Finalize v2017.3
Alan Baker [Wed, 10 Jan 2018 19:23:47 +0000 (14:23 -0500)]
Adding ostream operators for IR structures
* Added for Instruction, BasicBlock, Function and Module
* Uses new disassembly functionality that can disassemble individual
instructions
* For debug use only (no caching is done)
* Each output converts module to binary, parses and outputs an
individual instruction
* Added a test for whole module output
* Disabling Microsoft checked iterator warnings
* Updated check_copyright.py to accept 2018
Alan Baker [Thu, 11 Jan 2018 20:05:39 +0000 (15:05 -0500)]
Maintain instruction to block mapping in phi insertion
* Changed MemPass::InsertPhiInstructions to set basic blocks for new
phis
* Local SSA elim now maintains instr to block mapping
* Added a test and confirmed it fails without the updated phis
* IRContext::set_instr_block no longer builds the map if the analysis is
invalid
* Added instruction to block mapping verification to
IRContext::IsConsistent()
Greg Fischer [Fri, 29 Dec 2017 23:44:43 +0000 (16:44 -0700)]
InsertExtractElim: Optimize through VectorShuffle, Mix
This improves Extract replacement to continue through VectorShuffle.
It will also handle Mix with 0.0 or 1.0 in the a-value of the desired
component.
To facilitate optimization of VectorShuffle, the algorithm was refactored
to pass around the indices of the extract in a vector rather than pass the
extract instruction itself. This allows the indices to be modified as the
algorithm progresses.
Steven Perron [Tue, 9 Jan 2018 17:45:46 +0000 (12:45 -0500)]
Add generic folding function and use in CCP
The current folding routines have a very cumbersome interface, make them
harder to use, and not a obvious how to extend.
This change is to create a new interface for the folding routines, and
show how it can be used by calling it from CCP.
This does not make a significant change to the behaviour of CCP. In
general it should produce the same code as before; however it is
possible that an instruction that takes 32-bit integers as inputs and
the result is not a 32-bit integer or bool will not be folded as before.
Alan Baker [Mon, 18 Dec 2017 17:13:10 +0000 (12:13 -0500)]
Adding additional functionality to ADCE.
Modified ADCE to remove dead globals.
* Entry point and execution mode instructions are marked as alive
* Reachable functions and their parameters are marked as alive
* Instruction deletion now deferred until the end of the pass
* Eliminated dead insts set, added IsDead to calculate that value
instead
* Ported applicable dead variable elimination tests
* Ported dead constant elim tests
Added dead function elimination to ADCE
* ported dead function elim tests
Added handling of decoration groups in ADCE
* Uses a custom sorter to traverse decorations in a specific order
* Simplifies necessary checks
Updated -O and -Os pass lists.
Andrey Tuganov [Tue, 9 Jan 2018 16:08:44 +0000 (11:08 -0500)]
Fix validation rules for GLSL pack/unpack 2x32
Alan Baker [Thu, 4 Jan 2018 22:04:03 +0000 (17:04 -0500)]
Rewriting dead branch elimination.
Pass now paints live blocks and fixes constant branches and switches as
it goes. No longer requires structured control flow. It also removes
unreachable blocks as a side effect. It fixes the IR (phis) before doing
any code removal (other than terminator changes).
Added several unit tests for updated/new functionality.
Does not remove dead edge from a phi node:
* Checks that incoming edges are live in order to retain them
* Added BasicBlock::IsSuccessor
* added test
Fixing phi updates in the presence of extra backedge blocks
* Added tests to catch bug
Reworked how phis are updated
* Instead of creating a new Phi and RAUW'ing the old phi with it, I now
replace the phi operands, but maintain the def/use manager correctly.
For unreachable merge:
* When considering unreachable continue blocks the code now properly
checks whether the incoming edge will continue to be live.
Major refactoring for review
* Broke into 4 major functions
* marking live blocks
* marking structured targets
* fixing phis
* deleting blocks
Diego Novillo [Mon, 8 Jan 2018 16:56:57 +0000 (11:56 -0500)]
Fix constant propagation of induction variables.
This fixes https://github.com/KhronosGroup/SPIRV-Tools/issues/1143.
When an instruction transitions from constant to bottom (varying) in the
lattice, we were telling the propagator that the instruction was
varying, but never updating the actual value in the values table.
This led to incorrect value substitutions at the end of propagation.
The patch also re-enables CCP in -O and -Os.
David Neto [Mon, 8 Jan 2018 15:42:02 +0000 (10:42 -0500)]
Fix method comment for BasicBlock::MegeBlockIdIfAny
Fixes #1177
Lei Zhang [Sun, 7 Jan 2018 15:50:01 +0000 (10:50 -0500)]
Allow relaxing validation of pointers in logical addressing mode
In HLSL structured buffer legalization, pointer to pointer types
are emitted to indicate a structured buffer variable should be
treated as an alias of some other variable. We need an option to
relax the check of pointer types in logical addressing mode to
catch other validation errors.
Victor Lomuller [Thu, 21 Dec 2017 14:47:25 +0000 (14:47 +0000)]
Add loop descriptors and some required dominator tree extensions.
Add post-order tree iterator.
Add DominatorTreeNode extensions:
- Add begin/end methods to do pre-order and post-order tree traversal from a given DominatorTreeNode
Add DominatorTree extensions:
- Add begin/end methods to do pre-order and post-order tree traversal
- Tree traversal ignore by default the pseudo entry block
- Retrieve a DominatorTreeNode from a basic block
Add loop descriptor:
- Add a LoopDescriptor class to register all loops in a given function.
- Add a Loop class to describe a loop:
- Loop parent
- Nested loops
- Loop depth
- Loop header, merge, continue and preheader
- Basic blocks that belong to the loop
Correct a bug that forced dominator tree to be constantly rebuilt.
David Neto [Fri, 5 Jan 2018 23:08:31 +0000 (18:08 -0500)]
AnalyzeInstUse: Reuse the instruction lookup
David Neto [Fri, 5 Jan 2018 18:21:36 +0000 (13:21 -0500)]
Remove CCP from size and performance recipes, pending bugfixes
Currently CCP is incorrectly optimizing loops.
See https://github.com/KhronosGroup/SPIRV-Tools/issues/1143
Pierre Moreau [Wed, 3 Jan 2018 00:54:55 +0000 (01:54 +0100)]
Linker code cleanups
Turn `Linker::Link()` into free functions
As very little information was kept in the Linker class, we can get rid
of the whole class and have the `Link()` as free functions instead; the
environment target as well as the consumer are passed along through an
`spv_context` object.
The resulting linked_binary is passed as a pointer rather than a
reference to follow the Google C++ Style guidelines.
Addresses remaining comments from
https://github.com/KhronosGroup/SPIRV-Tools/pull/693 about the SPIR-V
linker.
Fix variable naming in the linker
Some of the variables were using mixed case, which did not follow the
Google C++ Style guidelines.
Linker: Use EXPECT_EQ when possible and update some test
* Replace occurrences of ASSERT_EQ by EXPECT_EQ when possible;
* Reformulated some of the error messages;
* Added the symbol name in the error message when there is a type or
decoration mismatch between the imported and exported declarations.
Opt: List all duplicates removed by RemoveDuplicatePass in the header
Opt: Make the const version of GetLabelInst() return a pointer
For consistency with the non-const version, as well as other similar
functions.
Opt: Rename function_end to EndInst()
As pointed out by dneto0 the previous name was quite confusing and could
be mistaken with a function returning an end iterator.
Also change the return type of the const version to a pointer rather
than a reference, for consistency.
Opt: Add performance comment to RemoveDuplicateTypes and decorations
This comment was requested during the review of
https://github.com/KhronosGroup/SPIRV-Tools/pull/693.
Opt: Add comments and fix variable naming in RemoveDuplicatePass
* Add missing comments to private functions;
* Rename variables that were using mixed case;
* Add TODO for moving AreTypesEqual out.
Linker: Remove commented out code and add TODOs
Linker: Merged together strings that were too much splitted
Implement a C++ RAII wrapper around spv_context
Steven Perron [Thu, 4 Jan 2018 16:37:19 +0000 (11:37 -0500)]
Allow getting the base pointer of an image load/store.
In value numbering, we treat loads and stores of images, ie OpImageLoad,
as a memory operation where it is interested in the "base address" of
the instruction. In those cases, it is an image instruction.
The problem is that `Instruction::GetBaseAddress()` does not account for
the image instructions, so the assert at the end to make sure it found
a valid base address for its addressing mode fails.
The solution is to look at the load/store instruction to determine how
the assertion should be done.
Fixes #1160.
Diego Novillo [Fri, 5 Jan 2018 14:34:18 +0000 (09:34 -0500)]
Fix infinite simulation cycles in SSA propagator.
This fixes https://github.com/KhronosGroup/SPIRV-Tools/issues/1159. I
had missed a nuance in the original algorithm. When simulating Phi
instructions, the SSA edges out of a Phi instruction should never be
added to the list of edges to simulate.
Phi instructions can be in SSA def-use cycles with other Phi
instructions. This was causing the propagator to fall into an infinite
loop when the same def-use edge kept being added to the queue.
The original algorithm in the paper specifically separates the visit of
a Phi instruction vs the visit of a regular instruction. This fix makes
the implementation match the original algorithm.
Pierre Moreau [Thu, 4 Jan 2018 22:16:39 +0000 (23:16 +0100)]
Ignore clang-format-diff.py from copyrights check
Having that script locally, for example to run check_code_format.sh
before sending a pull request, would cause the copyright notices check
to fail.
David Neto [Thu, 4 Jan 2018 23:52:38 +0000 (18:52 -0500)]
dead branch elim: Track killed backedges
When deleting branches and blocks, also remove them from
the backedges set, in case they were there.
This prevents us from keeping stale pointers to deleted Instruction
objects. That memory could be used later by another instruction,
incorrectly signaling that something has a backedge reference, and
the dead branch eliminator could end up deleting live blocks.
Adds accessor method ir::BasicBlock::terminator
Fixes https://github.com/KhronosGroup/SPIRV-Tools/issues/1168
David Neto [Thu, 4 Jan 2018 17:59:50 +0000 (12:59 -0500)]
Add --print-all optimizer option
Adds optimizer API to write disassembly to a given output stream
before each pass, and after the last pass.
Adds spirv-opt --print-all option to write disassembly to stderr
before each pass, and after the last pass.