summaryrefslogtreecommitdiffstats
path: root/llvm/lib/Analysis
Commit message (Collapse)AuthorAgeFilesLines
* Remove x86_sse42_crc32_64_8 intrinsic. It has no functional difference from ↵Craig Topper2013-10-151-1/+0
| | | | | | x86_sse42_crc32_32_8 and was not mapped to a clang builtin. I'm not even sure why this form of the instruction is even called out explicitly in the docs. Also add AutoUpgrade support to convert it into the other intrinsic with appropriate trunc and zext. llvm-svn: 192672
* Rename DataLayout variables TD -> DLMatt Arsenault2013-10-031-26/+25
| | | | llvm-svn: 191927
* CaptureTracking: Plug a loophole in the "too many uses" heuristic.Benjamin Kramer2013-10-031-0/+6
| | | | | | | | | | The heuristic was added to avoid spending too much compile time A specially crafted test case (PR17461, PR16474) with many uses on a select or bitcast instruction can still trigger the slow case. Add a check for that case. This only affects compile time, don't have a good way to test it. llvm-svn: 191896
* Remove the very substantial, largely unmaintained legacy PGOChandler Carruth2013-10-0212-3832/+0
| | | | | | | | | | | | | | | | | | | | infrastructure. This was essentially work toward PGO based on a design that had several flaws, partially dating from a time when LLVM had a different architecture, and with an effort to modernize it abandoned without being completed. Since then, it has bitrotted for several years further. The result is nearly unusable, and isn't helping any of the modern PGO efforts. Instead, it is getting in the way, adding confusion about PGO in LLVM and distracting everyone with maintenance on essentially dead code. Removing it paves the way for modern efforts around PGO. Among other effects, this removes the last of the runtime libraries from LLVM. Those are being developed in the separate 'compiler-rt' project now, with somewhat different licensing specifically more approriate for runtimes. llvm-svn: 191835
* Remove several unused variables.Rafael Espindola2013-10-011-1/+0
| | | | | | Patch by Alp Toker. llvm-svn: 191757
* SCEVExpander: Fix a regression I introduced by to eagerly adding RAII objects.Benjamin Kramer2013-10-011-1/+4
| | | | | | PR17425. llvm-svn: 191741
* Convert manual insert point restores to the new RAII object.Benjamin Kramer2013-09-301-47/+15
| | | | llvm-svn: 191675
* ObjectSizeOffsetEvaluator: Don't run into infinite recursion if we have a ↵Benjamin Kramer2013-09-291-5/+7
| | | | | | | | cyclic GEP. Those can occur in dead code. PR17402. llvm-svn: 191644
* TBAA: try to fix the dragonegg bots.Manman Ren2013-09-271-1/+3
| | | | llvm-svn: 191585
* Minor code simplificationMatt Arsenault2013-09-271-11/+8
| | | | llvm-svn: 191579
* Use type helper functionsMatt Arsenault2013-09-271-3/+2
| | | | llvm-svn: 191574
* TBAA: handle scalar TBAA format and struct-path aware TBAA format.Manman Ren2013-09-271-10/+81
| | | | | | | | | | | | | | | | Remove the command line argument "struct-path-tbaa" since we should not depend on command line argument to decide which format the IR file is using. Instead, we check the first operand of the tbaa tag node, if it is a MDNode, we treat it as struct-path aware TBAA format, otherwise, we treat it as scalar TBAA format. When clang starts to use struct-path aware TBAA format no matter whether struct-path-tbaa is no, and we can auto-upgrade existing bc files, the support for scalar TBAA format can be dropped. Existing testing cases are updated to use the struct-path aware TBAA format. llvm-svn: 191538
* MemoryBuiltins: Remove posix_memalign from the list and replace it with a TODO.Benjamin Kramer2013-09-241-1/+1
| | | | | | | This code isn't ready to deal with allocation functions where the return is not the allocated pointer. The checks below will reject posix_memalign anyways. llvm-svn: 191319
* MemoryBuiltins: Reinstate optimizing (uninitialized) loads from operator new.Benjamin Kramer2013-09-241-7/+7
| | | | llvm-svn: 191315
* MemoryBuiltins: Fix operator new bits.Benjamin Kramer2013-09-241-3/+3
| | | | | | We really don't want to optimize malloc return value checks away. llvm-svn: 191313
* Teach MemoryBuiltins and InstructionSimplify that operator new never returns ↵Benjamin Kramer2013-09-243-6/+20
| | | | | | | | | | | | NULL. This is safe per C++11 18.6.1.1p3: [operator new returns] a non-null pointer to suitably aligned storage (3.7.4), or else throw a bad_alloc exception. This requirement is binding on a replacement version of this function. Brings us a tiny bit closer to eliminating more vector push_backs. llvm-svn: 191310
* InstSimplify: Fold equality comparisons between non-inbounds GEPs.Benjamin Kramer2013-09-231-2/+15
| | | | | | | | | Overflow doesn't affect the correctness of equalities. Computing this is cheap, we just reuse the computation for the inbounds case and try to peel of more non-inbounds GEPs. This pattern is unlikely to ever appear in code generated by Clang, but SCEV occasionally produces it. llvm-svn: 191200
* Fix a constant folding address space place I missed.Matt Arsenault2013-09-171-3/+4
| | | | | | | | If address space 0 was smaller than the address space in a constant inttoptr/ptrtoint pair, the wrong mask size would be used. llvm-svn: 190899
* Move variable into assert to avoid unused variable warning.Eric Christopher2013-09-171-2/+1
| | | | llvm-svn: 190886
* Costmodel: Add support for horizontal vector reductionsArnold Schwaighofer2013-09-172-0/+281
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Upcoming SLP vectorization improvements will want to be able to estimate costs of horizontal reductions. Add infrastructure to support this. We model reductions as a series of (shufflevector,add) tuples ultimately followed by an extractelement. For example, for an add-reduction of <4 x float> we could generate the following sequence: (v0, v1, v2, v3) \ \ / / \ \ / + + (v0+v2, v1+v3, undef, undef) \ / ((v0+v2) + (v1+v3), undef, undef) %rdx.shuf = shufflevector <4 x float> %rdx, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef> %bin.rdx = fadd <4 x float> %rdx, %rdx.shuf %rdx.shuf7 = shufflevector <4 x float> %bin.rdx, <4 x float> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef> %bin.rdx8 = fadd <4 x float> %bin.rdx, %rdx.shuf7 %r = extractelement <4 x float> %bin.rdx8, i32 0 This commit adds a cost model interface "getReductionCost(Opcode, Ty, Pairwise)" that will allow clients to ask for the cost of such a reduction (as backends might generate more efficient code than the cost of the individual instructions summed up). This interface is excercised by the CostModel analysis pass which looks for reduction patterns like the one above - starting at extractelements - and if it sees a matching sequence will call the cost model interface. We will also support a second form of pairwise reduction that is well supported on common architectures (haddps, vpadd, faddp). (v0, v1, v2, v3) \ / \ / (v0+v1, v2+v3, undef, undef) \ / ((v0+v1)+(v2+v3), undef, undef, undef) %rdx.shuf.0.0 = shufflevector <4 x float> %rdx, <4 x float> undef, <4 x i32> <i32 0, i32 2 , i32 undef, i32 undef> %rdx.shuf.0.1 = shufflevector <4 x float> %rdx, <4 x float> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef> %bin.rdx.0 = fadd <4 x float> %rdx.shuf.0.0, %rdx.shuf.0.1 %rdx.shuf.1.0 = shufflevector <4 x float> %bin.rdx.0, <4 x float> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef> %rdx.shuf.1.1 = shufflevector <4 x float> %bin.rdx.0, <4 x float> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef> %bin.rdx.1 = fadd <4 x float> %rdx.shuf.1.0, %rdx.shuf.1.1 %r = extractelement <4 x float> %bin.rdx.1, i32 0 llvm-svn: 190876
* In AliasSetTracker, do not change the alias set to "mod/ref" when addingKrzysztof Parzyszek2013-09-121-2/+0
| | | | | | a volatile load, or a volatile store. llvm-svn: 190631
* Move variable under condition where it is usedMatt Arsenault2013-09-121-1/+2
| | | | llvm-svn: 190567
* Add getUnrollingPreferences to TTIHal Finkel2013-09-111-0/+7
| | | | | | | | | Allow targets to customize the default behavior of the generic loop unrolling transformation. This will be used by the PowerPC backend when targeting the A2 core (which is in-order with a deep pipeline), and using more aggressive defaults is important. llvm-svn: 190542
* Teach ScalarEvolution about pointer address spacesMatt Arsenault2013-09-102-12/+22
| | | | llvm-svn: 190425
* TBAA: add isTBAAVtableAccess to MDNode so clients can call the functionManman Ren2013-09-061-0/+19
| | | | | | | | | | | | | instead of having its own implementation. The implementation of isTBAAVtableAccess is in TypeBasedAliasAnalysis.cpp since it is related to the format of TBAA metadata. The path for struct-path tbaa will be exercised by test/Instrumentation/ThreadSanitizer/read_from_global.ll, vptr_read.ll, and vptr_update.ll when struct-path tbaa is on by default. llvm-svn: 190216
* Revert: r189565 - Add getUnrollingPreferences to TTIHal Finkel2013-08-291-9/+0
| | | | | | | | | | | | | | | Revert unintentional commit (of an unreviewed change). Original commit message: Add getUnrollingPreferences to TTI Allow targets to customize the default behavior of the generic loop unrolling transformation. This will be used by the PowerPC backend when targeting the A2 core (which is in-order with a deep pipeline), and using more aggressive defaults is important. llvm-svn: 189566
* Add getUnrollingPreferences to TTIHal Finkel2013-08-291-0/+9
| | | | | | | | | Allow targets to customize the default behavior of the generic loop unrolling transformation. This will be used by the PowerPC backend when targeting the A2 core (which is in-order with a deep pipeline), and using more aggressive defaults is important. llvm-svn: 189565
* Handle address spaces in TargetTransformInfoMatt Arsenault2013-08-281-7/+15
| | | | llvm-svn: 189527
* Fix lint assert on integer vector divisionMatt Arsenault2013-08-261-6/+34
| | | | llvm-svn: 189290
* Remove trailing spaces.Jakub Staszak2013-08-241-45/+44
| | | | llvm-svn: 189173
* Turn MipsOptimizeMathLibCalls into a target-independent scalar transformRichard Sandiford2013-08-231-0/+8
| | | | | | | | | | ...so that it can be used for z too. Most of the code is the same. The only real change is to use TargetTransformInfo to test when a sqrt instruction is available. The pass is opt-in because at the moment it only handles sqrt. llvm-svn: 189097
* Reorder headers according to lint.Bill Wendling2013-08-211-4/+2
| | | | llvm-svn: 188932
* Add some constantness.Jakub Staszak2013-08-201-5/+6
| | | | llvm-svn: 188844
* Teach ConstantFolding about pointer address spacesMatt Arsenault2013-08-201-33/+54
| | | | llvm-svn: 188831
* Fix assert with GEP ptr vector indexing structsMatt Arsenault2013-08-191-2/+12
| | | | | | | | Also fix it calculating the wrong value. The struct index is not a ConstantInt, so it was being interpreted as an array index. llvm-svn: 188713
* Fix SCEVExpander creating distinct duplicate PHI entriesHal Finkel2013-08-181-0/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | This fixes SCEVExpander so that it does not create multiple distinct induction variables for duplicate PHI entries. Specifically, given some code like this: do.body6: ; preds = %do.body6, %do.body6, %if.then5 %end.0 = phi i8* [ undef, %if.then5 ], [ %incdec.ptr, %do.body6 ], [ %incdec.ptr, %do.body6 ] ... Note that it is legal to have multiple entries for a basic block so long as the associated value is the same. So the above input is okay, but expanding an AddRec in this loop could produce code like this: do.body6: ; preds = %do.body6, %do.body6, %if.then5 %indvar = phi i64 [ %indvar.next, %do.body6 ], [ %indvar.next1, %do.body6 ], [ 0, %if.then5 ] %end.0 = phi i8* [ undef, %if.then5 ], [ %incdec.ptr, %do.body6 ], [ %incdec.ptr, %do.body6 ] ... %indvar.next = add i64 %indvar, 1 %indvar.next1 = add i64 %indvar, 1 And this is not legal because there are two PHI entries for %do.body6 each with a distinct value. Unfortunately, I don't have an in-tree test case. llvm-svn: 188614
* Fix an oversight in isPotentiallyReachable where we wouldn't do any CFG-walkingNick Lewycky2013-08-131-53/+70
| | | | | | | | | | | | to find loops if the From and To instructions were in the same block. Refactor the code a little now that we need to fill to start the CFG-walking algorithm with more than one starting basic block sometimes. Special thanks to Andrew Trick for catching an error in my understanding of natural loops in code review. llvm-svn: 188236
* Slightly simplify code with helper functionsMatt Arsenault2013-08-121-14/+16
| | | | | | | e.g. Use Ty->getPointerElementType() instead of cast<PointerType>(Ty)->getElementType() llvm-svn: 188223
* Add some braces, and spaces around operatorsMatt Arsenault2013-08-121-26/+42
| | | | llvm-svn: 188219
* Teach ValueTracking about address spacesMatt Arsenault2013-08-101-8/+12
| | | | llvm-svn: 188140
* Disable inlining between sanitized and non-sanitized functions.Evgeniy Stepanov2013-08-081-5/+23
| | | | | | | Inlining between functions with different values of sanitize_* attributes leads to over- or under-sanitizing, which is always bad. llvm-svn: 187967
* Add ISD::FROUND for libm round()Hal Finkel2013-08-071-0/+1
| | | | | | | | | | | | | | | All libm floating-point rounding functions, except for round(), had their own ISD nodes. Recent PowerPC cores have an instruction for round(), and so here I'm adding ISD::FROUND so that round() can be custom lowered as well. For the most part, this is straightforward. I've added an intrinsic and a matching ISD node just like those for nearbyint() and friends. The SelectionDAG pattern I've named frnd (because ISD::FP_ROUND has already claimed fround). This will be used by the PowerPC backend in a follow-up commit. llvm-svn: 187926
* Remove extraneous semicolon.Jakub Staszak2013-08-061-1/+1
| | | | llvm-svn: 187806
* Minor address space code simplification.Matt Arsenault2013-08-031-9/+3
| | | | | | Remove assertion that the verifier should catch. llvm-svn: 187692
* Teach InstructionSimplify about pointer address spacesMatt Arsenault2013-08-021-4/+9
| | | | llvm-svn: 187635
* Fix a severe compile time problem when forming large SCEV expressions.Andrew Trick2013-07-311-0/+3
| | | | | | | | | | | | This fix is very lightweight. The same fix already existed for AddRec but was missing for NAry expressions. This is obviously an improvement and I'm unsure how to test compile time problems. Patch by Xiaoyi Guo! llvm-svn: 187475
* isKnownToBeAPowerOfTwo: Strengthen isKnownToBeAPowerOfTwo's analysis on add ↵David Majnemer2013-07-301-16/+30
| | | | | | | | | instructions Call into ComputeMaskedBits to figure out which bits are set on both add operands and determine if the value is a power-of-two-or-zero or not. llvm-svn: 187445
* Also update CMakeLists.txt for r187283.Nick Lewycky2013-07-271-0/+1
| | | | llvm-svn: 187284
* Reimplement isPotentiallyReachable to make nocapture deduction much stronger.Nick Lewycky2013-07-272-22/+230
| | | | | | | | | | Adds unit tests for it too. Split BasicBlockUtils into an analysis-half and a transforms-half, and put the analysis bits into a new Analysis/CFG.{h,cpp}. Promote isPotentiallyReachable into llvm::isPotentiallyReachable and move it into Analysis/CFG. llvm-svn: 187283
* SimplifyCFG: Use parallel-and and parallel-or mode to consolidate branch ↵Tom Stellard2013-07-271-0/+6
| | | | | | | | | | | | | | conditions Merge consecutive if-regions if they contain identical statements. Both transformations reduce number of branches. The transformation is guarded by a target-hook, and is currently enabled only for +R600, but the correctness has been tested on X86 target using a variety of CPU benchmarks. Patch by: Mei Ye llvm-svn: 187278
OpenPOWER on IntegriCloud