aboutsummaryrefslogtreecommitdiff
path: root/lib/Transforms
AgeCommit message (Collapse)Author
2008-06-11For now, avoid generating FP select instructions in order to speculatively ↵Evan Cheng
execute integer arithmetic instructions. FP selects are more likely to be expensive (even compared to branch on fcmp). This is not a wonderful solution but I rather err on the side of conservative. This fixes the heapsort performance regressions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52224 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-11Avoid duplicating loop header which leads to unnatural loops (and just seem ↵Evan Cheng
like general badness to me, likely to cause code explosion). Patch by Florian Brandner. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52223 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-11Teach instruction combining about the extractvalue. It can succesfully foldMatthijs Kooijman
useless insert-extract chains, similar to how it folds them for vectors. Add a testcase for this. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52217 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-11Clarify a comment.Matthijs Kooijman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52212 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-10op_iterator-ify loopsGabor Greif
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52191 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-09lower calls to abs to inline code, PR2337Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52138 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-09Fix PR2411, where ip constant prop would propagate theChris Lattner
result of a weak function. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52137 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-08Remove comparison methods for MVT. The main causeDuncan Sands
of apint codegen failure is the DAG combiner doing the wrong thing because it was comparing MVT's using < rather than comparing the number of bits. Removing the < method makes this mistake impossible to commit. Instead, add helper methods for comparing bits and use them. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52098 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-08Limit the icmp+phi merging optimization to the cases where it is profitable:Chris Lattner
don't make i1 phis when it won't be possible to eliminate them. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52097 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-07Speculatively execute a block when the the block is the then part of a ↵Evan Cheng
triangle shape and it contains a single, side effect free, cheap instruction. The branch is eliminated by adding a select instruction. i.e. Turn BB: %t1 = icmp br i1 %t1, label %BB1, label %BB2 BB1: %t3 = add %t2, c br label BB2 BB2: => BB: %t1 = icmp %t4 = add %t2, c %t3 = select i1 %t1, %t2, %t3 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52073 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-06LoopSimplify preserves AA.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52053 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-06Wrap MVT::ValueType in a struct to get type safetyDuncan Sands
and better control the abstraction. Rename the type to MVT. To update out-of-tree patches, the main thing to do is to rename MVT::ValueType to MVT, and rewrite expressions like MVT::getSizeInBits(VT) in the form VT.getSizeInBits(). Use VT.getSimpleVT() to extract a MVT::SimpleValueType for use in switch statements (you will get an assert failure if VT is an extended value type - these shouldn't exist after type legalization). This results in a small speedup of codegen and no new testsuite failures (x86-64 linux). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52044 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-06As Chris suggested, handle the situation if ShAmt larger than BitWidth, Zhou Sheng
otherwise, opt might crash. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52041 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-05If BitWidth equals to ShtAmt, the RHSKnownZero[BitWidth-ShiftAmt-1] willZhou Sheng
crash the opt. Just fix this. Test case in llvm/test/Transforms/InstCombine/2008-06-05-ashr-crash.ll git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52003 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-05Learn ScalarReplAggregrates how stores and loads of first class aggregratesMatthijs Kooijman
work and how to replace them into individual values. Also, when trying to replace an aggregrate that is used by load or store with a single (large) integer, don't crash (but don't replace the aggregrate either). Also adds a testcase for both structs and arrays. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51997 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-05Let StructRetPromotion check if all if its users are really calls or invokesn,Matthijs Kooijman
not other instructions. This fixes a crash with the added testcase. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51992 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-05Let StructRetPromotion check if it's users are really calling it and notMatthijs Kooijman
passing its pointer. Fixes test with added testcase. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51991 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-05Use use_iterator::getOperandNo instead of CallSite::hasArgument to check if aMatthijs Kooijman
function is passed as an argument instead of called. Also do this check a bit earlier. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51990 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-05Update comments and documentation to reflect that GCSE and ValueNumbering areMatthijs Kooijman
deprecated by the GVN and GVNPRE passes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51983 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-04Remove unneeded #include.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51955 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-04Replace two manual loops with calls to CallSite::hasArguments (no functional ↵Matthijs Kooijman
changes). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51947 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-04Change packed struct layout so that field sizesDuncan Sands
are the same as in unpacked structs, only field positions differ. This only matters for structs containing x86 long double or an apint; it may cause backwards compatibility problems if someone has bitcode containing a packed struct with a field of one of those types. The issue is that only 10 bytes are needed to hold an x86 long double: the store size is 10 bytes, but the ABI size is 12 or 16 bytes (linux/ darwin) which comes from rounding the store size up by the alignment. Because it seemed silly not to pack an x86 long double into 10 bytes in a packed struct, this is what was done. I now think this was a mistake. Reserving the ABI size for an x86 long double field even in a packed struct makes things more uniform: the ABI size is now always used when reserving space for a type. This means that developers are less likely to make mistakes. It also makes life easier for the CBE which otherwise could not represent all LLVM packed structs (PR2402). Front-end people might need to adjust the way they create LLVM structs - see following change to llvm-gcc. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51928 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-03LoopIndexSplit can sometimes result in cases where a block in its own ↵Owen Anderson
domfrontier. Don't crash when we encounter one of these. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51915 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-03Fix whitespace in whitespace-significant pseudocode in a comment.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51890 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-02Update dom tree. Fix PR 2372.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51887 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-02move CannotBeNegativeZero to ValueTracking. Simplify some signbit comparisons.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51864 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-02move ComputeMaskedBits, MaskedValueIsZero, and ComputeNumSignBitsChris Lattner
out of instcombine into a new file in libanalysis. This also teaches ComputeNumSignBits about the number of sign bits in a constantint. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51863 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-01Fix two issues that Eli Friedman pointed out, where would misoptimized code ↵Owen Anderson
like: char a[200]; init(a, a+200); OR int a[200]; char* b = (char*)a; char* c = (char*)a; foo(b, c); git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51850 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-01Don't remove the memcpy when call slot substitution fails.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51848 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-01When simplifying a call to a bitcast function, tighten upDuncan Sands
the conditions for performing the transform when only the function declaration is available: no longer allow turning i32 into i64 for example. Only allow changing between pointer types, and between pointer types and integers of the same size. For return values ptr -> intptr was already allowed; I added ptr -> ptr and intptr -> ptr while there. As shown by a recent objc testcase, changing the way parameters/return values are passed can be fatal when calling code written in assembler that directly manipulates call arguments and return values unless the transform has no impact on the way they are passed at the codegen level. While it is possible to imagine an ABI that treats integers of pointer size differently to pointers, I don't think LLVM supports any so the transform should now be safe while still being useful. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51834 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-31Peer through sext/zext when looking for not(cmp).Nick Lewycky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51819 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-31Add more i1 optimizations. add, sub, mul, s/udiv on i1 are now simplified away.Nick Lewycky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51817 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-31Adding i1 is always Xor.Nick Lewycky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51816 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-30rewrite operand loops to use iteratorsGabor Greif
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51789 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-30Since LCSSA switched over to DenseMap, we have to be more careful to avoid ↵Owen Anderson
iterator invalidation. Fixes PR2385. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51777 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-30Use eraseFromParent() instead of doing that manually in two places.Matthijs Kooijman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51770 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-29const-ify getOpcode.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51698 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-29Add a newline at the end of this file.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51680 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-29Replace the old ADCE implementation with a new one that more simply solvesOwen Anderson
the one case that ADCE catches that normal DCE doesn't: non-induction variable loop computations. This implementation handles this problem without using postdominators. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51668 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-29Remove debugging code.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51666 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-29convert more operand loops to iterator formulationGabor Greif
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51663 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-28Implement PR2370: memmove(x,x,size) -> noop.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51636 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-27Fix some constructs that gcc-4.4 warns about.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51591 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-27InequalityGraph::node() can create new nodes, invalidating iterators acrossNick Lewycky
the set of nodes. Fix makeEqual to handle this by creating the new node first then iterating across them second. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51573 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-26Grammaro.Nick Lewycky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51572 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-26Factor code to copy global value attributes likeDuncan Sands
the section or the visibility from one global value to another: copyAttributesFrom. This is particularly useful for duplicating functions: previously this was done by explicitly copying each attribute in turn at each place where a new function was created out of an old one, with the result that obscure attributes were regularly forgotten (like the collector or the section). Hopefully now everything is uniform and nothing is forgotten. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51567 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-26Use a DenseMap instead of an std::map, speeding up the testcase in PR2368 by ↵Owen Anderson
about a third. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51565 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-25"ret (constexpr)" can't be folded into a Constant. Add a method toNick Lewycky
Analysis/ConstantFolding to fold ConstantExpr's, then make instcombine use it to try to use targetdata to fold constant expressions on void instructions. Also extend the icmp(inttoptr, inttoptr) folding to handle the case where int size != ptr size. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51559 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-24Fix a serious brain-o. Obviously no-one reviewed my patch :(Chris Lattner
This fixes PR2359 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51536 91177308-0d34-0410-b5e6-96231b3b80d8
2008-05-24Fix PR2358 by resolving calls with undef arguments to overdefined.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@51535 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-08-05 02:51:18 +0000