aboutsummaryrefslogtreecommitdiff
path: root/lib/Transforms
AgeCommit message (Collapse)Author
2013-05-14LoopVectorize: Handle loops with multiple forward inductionsArnold Schwaighofer
We used to give up if we saw two integer inductions. After this patch, we base further induction variables on the chosen one like we do in the reverse induction and pointer induction case. Fixes PR15720. radar://13851975 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181746 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-14[objc-arc-opts] Added debug statements when we set and unset whether a ↵Michael Gottesman
pointer is known positive. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181745 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-13[objc-arc-opts] In the presense of an alloca unconditionally remove RR pairs ↵Michael Gottesman
if and only if we are both KnownSafeBU/KnownSafeTD rather than just either or. In the presense of a block being initialized, the frontend will emit the objc_retain on the original pointer and the release on the pointer loaded from the alloca. The optimizer will through the provenance analysis realize that the two are related (albiet different), but since we only require KnownSafe in one direction, will match the inner retain on the original pointer with the guard release on the original pointer. This is fixed by ensuring that in the presense of allocas we only unconditionally remove pointers if both our retain and our release are KnownSafe (i.e. we are KnownSafe in both directions) since we must deal with the possibility that the frontend will emit what (to the optimizer) appears to be unbalanced retain/releases. An example of the miscompile is: %A = alloca retain(%x) retain(%x) <--- Inner Retain store %x, %A %y = load %A ... DO STUFF ... release(%y) call void @use(%x) release(%x) <--- Guarding Release getting optimized to: %A = alloca retain(%x) store %x, %A %y = load %A ... DO STUFF ... release(%y) call void @use(%x) rdar://13750319 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181743 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-13Move a couple more statistics inside '#ifndef NDEBUG'.Matt Beaumont-Gay
Suppresses an unused-variable warning in -Asserts builds. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181733 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-13[objc-arc-opts] Add comment to BBState making it clear that ↵Michael Gottesman
get{TopDown,BottomUp}PtrState will create a new PtrState object if it does not find a PtrState for Arg. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181726 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-13[objc-arc] Move the before optimization statistics gathering phase out of ↵Michael Gottesman
OptimizeIndividualCalls. This makes the statistics gathering completely independent of the actual optimization occuring, preventing any sort of bleeding over from occuring. Additionally, it simplifies a switch statement in the non-statistic gathering case. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181719 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-13Suppress GCC compiler warnings in release builds about variables that are onlyDuncan Sands
read in asserts. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181689 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-13SLPVectorizer: Swap LHS and RHS. No functionality change.Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181684 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-12SLPVectorizer: Fix a bug in the code that generates extracts for values with ↵Nadav Rotem
multiple users. The external user does not have to be in lane #0. We have to save the lane for each scalar so that we know which vector lane to extract. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181674 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-12SLPVectorizer: Clear the map that maps between scalars to vectors after each ↵Nadav Rotem
round of vectorization. Testcase in the next commit. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181673 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-12InstCombine: Flip the order of two urem transformsDavid Majnemer
There are two transforms in visitUrem that conflict with each other. *) One, if a divisor is a power of two, subtracts one from the divisor and turns it into a bitwise-and. *) The other unwraps both operands if they are surrounded by zext instructions. Flipping the order allows the subtraction to go beneath the sign extension. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181668 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-11LoopVectorize: Use the widest induction variable typeArnold Schwaighofer
Use the widest induction type encountered for the cannonical induction variable. We used to turn the following loop into an empty loop because we used i8 as induction variable type and truncated 1024 to 0 as trip count. int a[1024]; void fail() { int reverse_induction = 1023; unsigned char forward_induction = 0; while ((reverse_induction) >= 0) { forward_induction++; a[reverse_induction] = forward_induction; --reverse_induction; } } radar://13862901 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181667 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-11LoopVectorize: Use variable instead of repeated function callArnold Schwaighofer
No functionality change intended. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181666 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-11LoopVectorize: Use IRBuilder interface in more placesArnold Schwaighofer
No functionality change intended. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181665 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-11InstCombine: Turn urem to bitwise-and more oftenDavid Majnemer
Use isKnownToBeAPowerOfTwo in visitUrem so that we may more aggressively fold away urem instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181661 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-10SLPVectorizer: Add support for trees with external users.Nadav Rotem
For example: bar() { int a = A[i]; int b = A[i+1]; B[i] = a; B[i+1] = b; foo(a); <--- a is used outside the vectorized expression. } git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181648 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-10Add a debug printNadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181647 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-10InstCombine: Don't claim to be able to evaluate any shl in a zexted type.Benjamin Kramer
The shift amount may be larger than the type leading to undefined behavior. Limit the transform to constant shift amounts. While there update the bits to clear in the result which may enable additional optimizations. PR15959. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181604 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-10InstCombine: Verify the type before transforming uitofp into select.Benjamin Kramer
PR15952. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181586 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-09Fix a documentation warning: \bried -> \briefDmitri Gribenko
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181551 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-09[GVN] Split critical-edge on the fly, instead of postpone edge-splitting to nextShuxin Yang
iteration. This on step toward non-iterative GVN. My local hack suggests that getting rid of iteration will speedup GVN by 30%+ on a medium sized input (2k LOC, C++). I cannot explain why not 2x or more at this moment. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181532 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-09Don't replace an alias in llvm.used with its target.Rafael Espindola
When we replace an internal alias with its target, be careful not to replace the entry in llvm.used (and llvm.compiler_used). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181524 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-09InstCombine: Don't just copy known bits from the first operand of an srem.Benjamin Kramer
That's obviously wrong. Conservatively restrict it to the sign bit, which matches the original intention of this analysis. Fixes PR15940. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181518 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-09LoopVectorizer: Don't assert on the absence of induction variablesArnold Schwaighofer
A computable loop exit count does not imply the presence of an induction variable. Scalar evolution can return a value for an infinite loop. Fixes PR15926. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181495 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-08Add DebugIR pass -- emits IR file and replace source lines with IR lines in MDDaniel Malea
- requires existing debug information to be present - fixes up file name and line number information in metadata - emits a "<orig_filename>-debug.ll" succinct IR file (without !dbg metadata or debug intrinsics) that can be read by a debugger - initialize pass in opt tool to enable the "-debug-ir" flag - lit tests to follow git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181467 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-08Fix a bug in codegenprep where it was losing track of values OptimizeMemoryInstNick Lewycky
by switching to a ValueMap. Patch by Andrea DiBiagio! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181397 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-07LoopVectorizer: Improve reduction variable identificationArnold Schwaighofer
The two nested loops were confusing and also conservative in identifying reduction variables. This patch replaces them by a worklist based approach. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181369 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-07LoopVectorize: getConsecutiveVector must respect signed arithmeticArnold Schwaighofer
We were passing an i32 to ConstantInt::get where an i64 was needed and we must also pass the sign if we pass negatives numbers. The start index passed to getConsecutiveVector must also be signed. Should fix PR15882. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181286 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-06InstCombine: (X ^ signbit) + C -> X + (signbit ^ C)David Majnemer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181249 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-06Rotate multi-exit loops even if the latch was simplified.Andrew Trick
Test case by Michele Scandale! Fixes PR10293: Load not hoisted out of loop with multiple exits. There are few regressions with this patch, now tracked by rdar:13817079, and a roughly equal number of improvements. The regressions are almost certainly back luck because LoopRotate has very little idea of whether rotation is profitable. Doing better requires a more comprehensive solution. This checkin is a quick fix that lacks generality (PR10293 has a counter-example). But it trivially fixes the case in PR10293 without interfering with other cases, and it does satify the criteria that LoopRotate is a loop canonicalization pass that should avoid heuristics and special cases. I can think of two approaches that would probably be better in the long run. Ultimately they may both make sense. (1) LoopRotate should check that the current header would make a good loop guard, and that the loop does not already has a sufficient guard. The artifical SimplifiedLoopLatch check would be unnecessary, and the design would be more general and canonical. Two difficulties: - We need a strong guarantee that we won't endlessly rotate, so the analysis would need to be precise in order to avoid the SimplifiedLoopLatch precondition. - Analysis like this are usually based on SCEV, which we don't want to rely on. (2) Rotate on-demand in late loop passes. This could even be done by shoving the loop back on the queue after the optimization that needs it. This could work well when we find LICM opportunities in multi-branch loops. This requires some work, and it doesn't really solve the problem of SCEV wanting a loop guard before the analysis. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181230 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-06Provide InstCombines for the following 3 cases:Jean-Luc Duprat
A * (1 - (uitofp i1 C)) -> select C, 0, A B * (uitofp i1 C) -> select C, B, 0 select C, 0, A + select C, B, 0 -> select C, B, A These come up in code that has been hand-optimized from a select to a linear blend, on platforms where that may have mattered. We want to undo such changes with the following transform: A*(1 - uitofp i1 C) + B*(uitofp i1 C) -> select C, A, B git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181216 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-06Update the comment to mention that we use TTI.Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181178 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-06Revert r164763 because it introduces new shuffles.Nadav Rotem
Thanks Nick Lewycky for pointing this out. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181177 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-06Fix const merging when an alias of a const is llvm.used.Rafael Espindola
We used to disable constant merging not only if a constant is llvm.used, but also if an alias of a constant is llvm.used. This change fixes that. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181175 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-05LoopVectorize: Print values instead of pointers in debug output.Benjamin Kramer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181157 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-05LoopVectorize: Add support for floating point min/max reductionsArnold Schwaighofer
Add support for min/max reductions when "no-nans-float-math" is enabled. This allows us to assume we have ordered floating point math and treat ordered and unordered predicates equally. radar://13723044 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181144 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-05LoopVectorizer: Cleanup of miminimum/maximum pattern match codeArnold Schwaighofer
No need for setting the operands. The pointers are going to be bound by the matcher. radar://13723044 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181142 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-05LoopVectorize: We don't need an identity element for min/max reductionsArnold Schwaighofer
We can just use the initial element that feeds the reduction. max(max(x, y), z) == max(max(x,y), max(x,z)) radar://13723044 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181141 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-05Add ArrayRef constructor from None, and do the cleanups that this ↵Dmitri Gribenko
constructor enables Patch by Robert Wilhelm. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181138 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-04Tabs to spaces. No functionality change.Nick Lewycky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181082 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-03Decompose GVN::processNonLocalLoad() (about 400 LOC) into smaller helper ↵Shuxin Yang
functions. No function change. This function consists of following steps: 1. Collect dependent memory accesses. 2. Analyze availability. 3. Perform fully redundancy elimination, or 4. Perform PRE, depending on the availability Step 2, 3 and 4 are now moved to three helper routines. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181047 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-03LoopVectorizer: Add support for if-conversion of PHINodes with 3+ incoming ↵Nadav Rotem
values. By supporting the vectorization of PHINodes with more than two incoming values we can increase the complexity of nested if statements. We can now vectorize this loop: int foo(int *A, int *B, int n) { for (int i=0; i < n; i++) { int x = 9; if (A[i] > B[i]) { if (A[i] > 19) { x = 3; } else if (B[i] < 4 ) { x = 4; } else { x = 5; } } A[i] = x; } } git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181037 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02[GV] Remove dead code which is really difficult to decipher.Shuxin Yang
Actually it took me couple of hours trying to make sense of them and only to find they are dead code. I guess the original author used "allSingleSucc" to indicate if there are any critial edge emanating from some blocks, and tried to perform code motion (actually speculation) in the presence of these critical edges; but later on he/she changed mind and decided to perform edge-splitting first. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180951 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01This patch breaks up Wrap.h so that it does not have to include all of Filip Pizlo
the things, and renames it to CBindingWrapping.h. I also moved CBindingWrapping.h into Support/. This new file just contains the macros for defining different wrap/unwrap methods. The calls to those macros, as well as any custom wrap/unwrap definitions (like for array of Values for example), are put into corresponding C++ headers. Doing this required some #include surgery, since some .cpp files relied on the fact that including Wrap.h implicitly caused the inclusion of a bunch of other things. This also now means that the C++ headers will include their corresponding C API headers; for example Value.h must include llvm-c/Core.h. I think this is harmless, since the C API headers contain just external function declarations and some C types, so I don't believe there should be any nasty dependency issues here. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180881 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01SROA: Generate selects instead of shuffles when blending values because this ↵Nadav Rotem
is the cannonical form. Shuffles are more difficult to lower and we usually don't touch them, while we do optimize selects more often. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180875 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01Revert "InstCombine: Fold more shuffles of shuffles."Jim Grosbach
This reverts commit r180802 There's ongoing discussion about whether this is the right place to make this transformation. Reverting for now while we figure it out. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180834 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-30Fix a use after free. RI is freed before the call to getDebugLoc(). ToRichard Trieu
prevent this, capture the location before RI is freed. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180824 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-30Fix a typoNadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180806 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-30InstCombine: Fold more shuffles of shuffles.Jim Grosbach
Always fold a shuffle-of-shuffle into a single shuffle when there's only one input vector in the first place. Continue to be more conservative when there's multiple inputs. rdar://13402653 PR15866 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180802 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-30Spelling. Thanks, Eric.Adrian Prantl
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180794 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-04-27 22:43:57 +0000