summaryrefslogtreecommitdiffstats
path: root/llvm/lib/Transforms
Commit message (Collapse)AuthorAgeFilesLines
* LSR critical edge splitting fix for PR13756.Andrew Trick2012-09-181-11/+15
| | | | llvm-svn: 164147
* Fix getCommonType in a different way from the way I fixed it whenChandler Carruth2012-09-181-1/+1
| | | | | | | | | | | | working on FCA splitting. Instead of refusing to form a common type when there are uses of a subsection of the alloca as well as a use of the entire alloca, just skip the subsection uses and continue looking for a whole-alloca use with a type that we can use. This produces slightly prettier IR I think, and also fixes the other failure in the test. llvm-svn: 164146
* Fix build for compilers that don't understand injected class names properly.Benjamin Kramer2012-09-181-2/+2
| | | | llvm-svn: 164142
* SROA: Use CRTP for OpSplitter to get rid of virtual dispatch and the ↵Benjamin Kramer2012-09-181-8/+7
| | | | | | virtual-dtor warnings that come with it. llvm-svn: 164140
* SROA: Replace the member function template contraption for recursively ↵Benjamin Kramer2012-09-181-97/+105
| | | | | | | | splitting aggregates into a real class. No intended functionality change. llvm-svn: 164135
* SROA.cpp: Appease msvc.NAKAMURA Takumi2012-09-181-1/+1
| | | | | | ...I don't know why this could appease msvc...baad. llvm-svn: 164130
* LNT builders have picked up new SROA, disable it to get the remaining ↵Benjamin Kramer2012-09-181-1/+1
| | | | | | builders green again. llvm-svn: 164124
* Fix a warning in release builds and a test case I forgot to update withChandler Carruth2012-09-181-0/+1
| | | | | | a fix to getCommonType in the previous patch. llvm-svn: 164120
* Add a major missing piece to the new SROA pass: aggressive splitting ofChandler Carruth2012-09-182-8/+221
| | | | | | | | | | | | | | | | | | | | | | | FCAs. This is essential in order to promote allocas that are used in struct returns by frontends like Clang. The FCA load would block the rest of the pass from firing, resulting is significant regressions with the bullet benchmark in the nightly test suite. Thanks to Duncan for repeated discussions about how best to do this, and to both him and Benjamin for review. This appears to have blocked many places where the pass tries to fire, and so I'm expect somewhat different results with this fix added. As with the last big patch, I'm including a change to enable the SROA by default *temporarily*. Ben is going to remove this as soon as the LNT bots pick up the patch. I'm just trying to get a round of LNT numbers from the stable machines in the lab. NOTE: Four clang tests are expected to fail in the brief window where this is enabled. Sorry for the noise! llvm-svn: 164119
* Fix instcombine to obey requested alignment when merging allocas.Richard Osborne2012-09-181-3/+7
| | | | llvm-svn: 164117
* Mark unimplemented copy constructors and copy assignment operators as ↵Craig Topper2012-09-181-3/+2
| | | | | | LLVM_DELETED_FUNCTION. llvm-svn: 164090
* PGO: preserve branch-weight metadata when simplifying Switch to a sub, an icmpManman Ren2012-09-181-1/+38
| | | | | | and a conditional branch; also when removing dead cases from a switch. llvm-svn: 164084
* PGO: preserve branch-weight metadata when simplifying SwitchManman Ren2012-09-171-0/+15
| | | | | | | Hanlde the case when we split the default edge if the default target has "icmp" and unconditinal branch. llvm-svn: 164076
* PGO: preserve branch-weight metadata when simplifying SwitchOnSelect.Manman Ren2012-09-171-5/+28
| | | | llvm-svn: 164068
* PGO: preserve branch-weight metadata when simplifying two branches with a commonManman Ren2012-09-171-0/+27
| | | | | | destination in SimplifyCondBranchToCondBranch. llvm-svn: 164054
* s/__llvm_gcov_flush/__gcov_flush/gBill Wendling2012-09-171-4/+4
| | | | llvm-svn: 164040
* NewSROA: Provide a full set of operator< for ByteRanges.Benjamin Kramer2012-09-171-2/+7
| | | | | | MSVC8 won't compile lower_bound if one is missing. llvm-svn: 164035
* Fix a few vars that can end up being used without initialization.Axel Naumann2012-09-172-2/+2
| | | | | | The cases where no initialization happens should still be checked for logic flaws. llvm-svn: 164032
* Refactor the SROA visitors for partitioning an alloca and buildingChandler Carruth2012-09-161-24/+29
| | | | | | | | | | | | | | | | | | | | | partition use lists a bit. No functionality changed. These visitors are actually visiting a tuple of a Use and an offset into the alloca. However, we use the InstVisitor to handle the dispatch over the users, and so the Use and Offset are stored in class member variables and set just before each call to visit(). This is fairly awkward and makes the functions a bit harder to read, but its the only real option we have until InstVisitor can be rewritten to use variadic templates. However, this pattern shouldn't be followed on the helper member functions where there is no interface constraint from the visitor. We already were passing the instruction as a normal parameter rather than use the Use to get at it, start passing the offset as well. This will become more important in subsequent patches as the offset will in some cases change while visiting a single instruction. llvm-svn: 164003
* Use LLVM_DELETED_FUNCTION in place of 'DO NOT IMPLEMENT' comments.Craig Topper2012-09-152-4/+6
| | | | llvm-svn: 163974
* Disable new sroa now that all buildbots have tested it.Benjamin Kramer2012-09-151-1/+1
| | | | | | | | | | | | | | | | | What we have so far: - Some clang test failures (these were known already) - Perf results are mixed, some big regressions http://llvm.org/perf/db_default/v4/nts/3844 http://llvm.org/perf/db_default/v4/nts/3845 bullet suffers a lot. matmul is interesting: slower scalar code, faster with -vectorize. - Some dragonegg selfhost bots crash in SROA during selfhost now http://lab.llvm.org:8011/builders/dragonegg-x86_64-linux-gcc-4.6-self-host-checks/builds/1632 http://lab.llvm.org:8011/builders/dragonegg-x86_64-linux-gcc-4.5-self-host/builds/1891 llvm-svn: 163968
* Port the SSAUpdater-based promotion logic from the old SROA pass to theChandler Carruth2012-09-152-14/+167
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | new one, and add support for running the new pass in that mode and in that slot of the pass manager. With this the new pass can completely replace the old one within the pipeline. The strategy for enabling or disabling the SSAUpdater logic is to do it by making the requirement of the domtree analysis optional. By default, it is required and we get the standard mem2reg approach. This is usually the desired strategy when run in stand-alone situations. Within the CGSCC pass manager, we disable requiring of the domtree analysis and consequentially trigger fallback to the SSAUpdater promotion. In theory this would allow the pass to re-use a domtree if one happened to be available even when run in a mode that doesn't require it. In practice, it lets us have a single pass rather than two which was simpler for me to wrap my head around. There is a hidden flag to force the use of the SSAUpdater code path for the purpose of testing. The primary testing strategy is just to run the existing tests through that path. One notable difference is that it has custom code to handle lifetime markers, and one of the tests has been enhanced to exercise that code. This has survived a bootstrap and the test suite without serious correctness issues, however my run of the test suite produced *very* alarming performance numbers. I don't entirely understand or trust them though, so more investigation is on-going. To aid my understanding of the performance impact of the new SROA now that it runs throughout the optimization pipeline, I'm enabling it by default in this commit, and will disable it again once the LNT bots have picked up one iteration with it. I want to get those bots (which are much more stable) to evaluate the impact of the change before I jump to any conclusions. NOTE: Several Clang tests will fail because they run -O3 and check the result's order of output. They'll go back to passing once I disable it again. llvm-svn: 163965
* PGO: preserve branch-weight metadata when simplifying two branches with a commonManman Ren2012-09-151-111/+42
| | | | | | | | | | | | | | | | destination. Updated previous implementation to fix a case not covered: // PBI: br i1 %x, TrueDest, BB // BI: br i1 %y, TrueDest, FalseDest The other case was handled correctly. // PBI: br i1 %x, BB, FalseDest // BI: br i1 %y, TrueDest, FalseDest Also tried to use 64-bit arithmetic instead of APInt with scale to simplify the computation. Let me know if you have other opinions about this. llvm-svn: 163954
* Remove comment.Bill Wendling2012-09-141-1/+1
| | | | llvm-svn: 163945
* PGO: preserve branch-weight metadata when simplifying a switch with a singleManman Ren2012-09-142-2/+33
| | | | | | case to a conditional branch and when removing dead cases. llvm-svn: 163942
* Stylistic and 80-col fixesEvan Cheng2012-09-142-8/+4
| | | | llvm-svn: 163940
* Review feedback from Duncan Sands. Alphabetize includes and simplifyAlex Rosenberg2012-09-141-3/+3
| | | | | | lit config. llvm-svn: 163928
* Try to fix the bots by detecting inconsistant branch-weight metadata.Manman Ren2012-09-141-4/+10
| | | | llvm-svn: 163926
* PGO: preserve branch-weight metadata when merging two switches whereManman Ren2012-09-141-5/+12
| | | | | | | the default target of the first switch is not the basic block the second switch is in (PredDefault != BB). llvm-svn: 163916
* Fix Doxygen issues:Dmitri Gribenko2012-09-141-4/+4
| | | | | | | | | | * wrap code blocks in \code ... \endcode; * refer to parameter names in paragraphs correctly (\arg is not what most people want -- it starts a new paragraph); * use \param instead of \arg to document parameters in order to be consistent with the rest of the codebase. llvm-svn: 163902
* SROA: Silence unused variable warnings in Release builds.Benjamin Kramer2012-09-141-1/+8
| | | | | | The NDEBUG hack is ugly, but I see no better solution. llvm-svn: 163900
* Rework the computation of a sub-structure natural type. There wereChandler Carruth2012-09-141-10/+20
| | | | | | | | | | | | | pointless checks in here, bad asserts, and just confusing code. I've also added a bit more to the comment to clarify what this function is really trying to do as it was not obvious to Duncan when studying it. Thanks to Duncan for helping me dig through the issue. No real functionality changed here in practical cases, and certainly no test case. This is just cleanup spotted by inspection. llvm-svn: 163897
* Rely on the recursive check for pointer types rather than adding anChandler Carruth2012-09-141-3/+0
| | | | | | | explicit check before recursing. A simplification requested by Duncan during review. llvm-svn: 163896
* Be a bit more aggressive in bailing out of this routine. Spotted byChandler Carruth2012-09-141-1/+1
| | | | | | | inspection by Duncan during review. My suspicion is that we would still have returned 0 anyways in this case, but doing it sooner is better. llvm-svn: 163895
* Add some comments clarifying that the GEP analysis for vector GEPs isChandler Carruth2012-09-141-1/+4
| | | | | | | | deeply suspicious and likely to go away eventually. Also fix a bogus comment about one of the checks in the vector GEP analysis. Based on review from Duncan. llvm-svn: 163894
* Move an instance variable to a local variable based on review by Duncan.Chandler Carruth2012-09-141-9/+16
| | | | | | | | Originally I had anticipated needing to thread this through more bits of the SROA pass itself, but that ended up not happening. In the end, this is a much simpler way to manange the variable. llvm-svn: 163893
* Add a comment about debug intrinsics that I *really* don't want toChandler Carruth2012-09-141-0/+2
| | | | | | forget from Duncan's review as a FIXME. llvm-svn: 163892
* Add two asserts that Duncan thought would help ensure things don't rotChandler Carruth2012-09-141-0/+2
| | | | | | unexpectedly in the future. More fixes from his code review. llvm-svn: 163891
* Actually keep the flag default-off for now. =/ That's what I get forChandler Carruth2012-09-141-1/+1
| | | | | | being busy testing this... llvm-svn: 163890
* Remove some dead, commented out code Duncan spotted in review.Chandler Carruth2012-09-141-4/+0
| | | | llvm-svn: 163889
* Wrap the dumping and printing routines in NDEBUG and LLVM_ENABLE_DUMP macros.Chandler Carruth2012-09-141-0/+6
| | | | llvm-svn: 163888
* Lots of comment fixes and cleanups from Duncan's review.Chandler Carruth2012-09-141-10/+12
| | | | llvm-svn: 163887
* SROA.cpp: Unbreak gcc, sorry!NAKAMURA Takumi2012-09-141-2/+2
| | | | llvm-svn: 163886
* SROA.cpp: Appease msvc. LLVM_ATTRIBUTE(s) should come front of "const".NAKAMURA Takumi2012-09-141-2/+2
| | | | llvm-svn: 163885
* Speculative change to try to fix older GCC versions that can't handleChandler Carruth2012-09-141-2/+2
| | | | | | the injected class name of a dependent base class here. llvm-svn: 163884
* Introduce a new SROA implementation.Chandler Carruth2012-09-144-2/+2642
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This is essentially a ground up re-think of the SROA pass in LLVM. It was initially inspired by a few problems with the existing pass: - It is subject to the bane of my existence in optimizations: arbitrary thresholds. - It is overly conservative about which constructs can be split and promoted. - The vector value replacement aspect is separated from the splitting logic, missing many opportunities where splitting and vector value formation can work together. - The splitting is entirely based around the underlying type of the alloca, despite this type often having little to do with the reality of how that memory is used. This is especially prevelant with unions and base classes where we tail-pack derived members. - When splitting fails (often due to the thresholds), the vector value replacement (again because it is separate) can kick in for preposterous cases where we simply should have split the value. This results in forming i1024 and i2048 integer "bit vectors" that tremendously slow down subsequnet IR optimizations (due to large APInts) and impede the backend's lowering. The new design takes an approach that fundamentally is not susceptible to many of these problems. It is the result of a discusison between myself and Duncan Sands over IRC about how to premptively avoid these types of problems and how to do SROA in a more principled way. Since then, it has evolved and grown, but this remains an important aspect: it fixes real world problems with the SROA process today. First, the transform of SROA actually has little to do with replacement. It has more to do with splitting. The goal is to take an aggregate alloca and form a composition of scalar allocas which can replace it and will be most suitable to the eventual replacement by scalar SSA values. The actual replacement is performed by mem2reg (and in the future SSAUpdater). The splitting is divided into four phases. The first phase is an analysis of the uses of the alloca. This phase recursively walks uses, building up a dense datastructure representing the ranges of the alloca's memory actually used and checking for uses which inhibit any aspects of the transform such as the escape of a pointer. Once we have a mapping of the ranges of the alloca used by individual operations, we compute a partitioning of the used ranges. Some uses are inherently splittable (such as memcpy and memset), while scalar uses are not splittable. The goal is to build a partitioning that has the minimum number of splits while placing each unsplittable use in its own partition. Overlapping unsplittable uses belong to the same partition. This is the target split of the aggregate alloca, and it maximizes the number of scalar accesses which become accesses to their own alloca and candidates for promotion. Third, we re-walk the uses of the alloca and assign each specific memory access to all the partitions touched so that we have dense use-lists for each partition. Finally, we build a new, smaller alloca for each partition and rewrite each use of that partition to use the new alloca. During this phase the pass will also work very hard to transform uses of an alloca into a form suitable for promotion, including forming vector operations, speculating loads throguh PHI nodes and selects, etc. After splitting is complete, each newly refined alloca that is a candidate for promotion to a scalar SSA value is run through mem2reg. There are lots of reasonably detailed comments in the source code about the design and algorithms, and I'm going to be trying to improve them in subsequent commits to ensure this is well documented, as the new pass is in many ways more complex than the old one. Some of this is still a WIP, but the current state is reasonbly stable. It has passed bootstrap, the nightly test suite, and Duncan has run it successfully through the ACATS and DragonEgg test suites. That said, it remains behind a default-off flag until the last few pieces are in place, and full testing can be done. Specific areas I'm looking at next: - Improved comments and some code cleanup from reviews. - SSAUpdater and enabling this pass inside the CGSCC pass manager. - Some datastructure tuning and compile-time measurements. - More aggressive FCA splitting and vector formation. Many thanks to Duncan Sands for the thorough final review, as well as Benjamin Kramer for lots of review during the process of writing this pass, and Daniel Berlin for reviewing the data structures and algorithms and general theory of the pass. Also, several other people on IRC, over lunch tables, etc for lots of feedback and advice. llvm-svn: 163883
* Handle the new !tbaa.struct metadata tags when converting a memcpy into scalarDan Gohman2012-09-131-0/+17
| | | | | | loads and stores. llvm-svn: 163844
* Extract code for reducing a type to a single value type into a helper function.Dan Gohman2012-09-131-15/+21
| | | | llvm-svn: 163817
* MemCpyOpt: When forming a memset from stores also take GEP constexprs into ↵Benjamin Kramer2012-09-131-3/+3
| | | | | | | | account. This is common when storing to global variables. llvm-svn: 163809
* Fix an 80 char line limit.Nadav Rotem2012-09-131-1/+2
| | | | llvm-svn: 163808
OpenPOWER on IntegriCloud