aboutsummaryrefslogtreecommitdiff
path: root/lib/Transforms
AgeCommit message (Collapse)Author
2010-09-10This transform is also performed by InstructionSimplify, remove the duplicate.Benjamin Kramer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113608 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-10Lower the unrolling theshold to 150. Empirical tests indicate that this is ↵Owen Anderson
a sweet spot in the performance per code size increase curve. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113595 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-09What the loop unroller cares about, rather than just not unrolling loops ↵Owen Anderson
with calls, is not unrolling loops that contain calls that would be better off getting inlined. This mostly comes up when an interleaved devirtualization pass has devirtualized a call which the inliner will inline on a future pass. Thus, rather than blocking all loops containing calls, add a metric for "inline candidate calls" and block loops containing those instead. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113535 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-09Revert r113439, which relaxed the requirement that loops containing calls ↵Owen Anderson
cannot be unrolled. After some discussion, there seems to be a better way to achieve the same effect. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113528 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-09r113526 introduced an unintended change to the loop unrolling threshold. ↵Owen Anderson
Revert it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113527 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-09Fix typo in code to cap the loop code size reduction calculation.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113526 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-09Use code-size reduction metrics to estimate the amount of savings we'll get ↵Owen Anderson
when we unroll a loop. Next step is to recalculate the threshold values given this new heuristic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113525 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-08Relax the "don't unroll loops containing calls" rule. Instead, when a loop ↵Owen Anderson
contains a call, lower the unrolling threshold to the optimize-for-size threshold. Basically, for loops containing calls, unrolling can still be profitable as long as the loop is REALLY small. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113439 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-08Generalize instcombine's support for combining multiple bit checks into a ↵Owen Anderson
single test. Patch by Dirk Steinke! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113423 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-07Add a separate unrolling threshold when the current function is being ↵Owen Anderson
optimized for size. The threshold value of 50 is arbitrary, and I chose it simply by analogy to the inlining thresholds, where the baseline unrolling threshold is slightly smaller than the baseline inlining threshold. This could undoubtedly use some tuning. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113306 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-07Fix a serious performance regression introduced by r108687 on linux:Chris Lattner
turning (fptrunc (sqrt (fpext x))) -> (sqrtf x) is great, but we have to delete the original sqrt as well. Not doing so causes us to do two sqrt's when building with -fmath-errno (the default on linux). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113260 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-07Fix major bug in thunk detection. Also verify the calling convention.Nick Lewycky
Switch from isWeakForLinker to mayBeOverridden which is more accurate. Add more statistics and debugging info. Add comments. Move static function outside anonymous namespace. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113190 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-06fix PR8067, an over-aggressive assertion in LICM.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113146 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-06Teach loop rotate to hoist trivially invariant instructionsChris Lattner
in the duplicated block instead of duplicating them. Duplicating them into the end of the loop and the preheader means that we got a phi node in the header of the loop, which prevented LICM from hoisting them. GVN would usually come around later and merge the duplicated instructions so we'd get reasonable output... except that anything dependent on the shoulda-been-hoisted value can't be hoisted. In PR5319 (which this fixes), a memory value didn't get promoted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113134 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-06pull a simple method out of LICM into a new Chris Lattner
Loop::hasLoopInvariantOperands method. Remove a useless and confusing Loop::isLoopInvariant(Instruction) method, which didn't do what you thought it did. No functionality change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113133 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-05more cleanupsChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113115 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-05Change lower atomic pass to use IntrinsicInst to simplify it a bit.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113114 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-05eliminate some non-obvious casts. UndefValue isa Constant.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113113 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-05Fix warning reported by MSVC++ builder.Nick Lewycky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113106 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-05Switch FnSet to containing the ComparableFunction instead of a pointer to one.Nick Lewycky
This reduces malloc traffic (yay!) and removes MergeFunctionsEqualityInfo. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113105 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-05Fix many bugs when merging weak-strong and weak-weak pairs. We now merge allNick Lewycky
strong functions first to make sure they're the canonical definitions and then do a second pass looking only for weak functions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113104 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-04zap dead code.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113073 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-04Fix LoopSimplify to notify ScalarEvolution when splitting a loop backedgeDan Gohman
into an inner loop, as the new loop iteration may differ substantially. This fixes PR8078. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113057 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-04fix a bug in my licm rewrite when a load from the promoted memoryChris Lattner
location is being re-stored to the memory location. We would get a dangling pointer from the SSAUpdate data structure and miss a use. This fixes PR8068 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113042 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-03Propagate non-local comparisons. Fixes PR1757.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113025 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-03Add support for simplifying a load from a computed value to a load from a ↵Owen Anderson
global when it is provable that they're equivalent. This fixes PR4855. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112994 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-02fix more AST updating bugs, correcting miscompilation in PR8041Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112878 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-02Reapply commit 112699, speculatively reverted by echristo, sinceDuncan Sands
I'm sure it is harmless. Original commit message: If PrototypeValue is erased in the middle of using the SSAUpdator then the SSAUpdator may access freed memory. Instead, simply pass in the type and name explicitly, which is all that was used anyway. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112810 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-01deepen my MMX/SRoA hack to avoid hurting non-x86 codegen.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112763 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-01Fix loop unswitching's assumption that a code path which eitherDan Gohman
infinite loops or exits will eventually exit. This fixes PR5373. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112745 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-01JumpThreading keeps LazyValueInfo up to date, so we don't need to rerun itOwen Anderson
if we schedule another LVI-using pass afterwards. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112722 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-01Speculatively revert 112699 and 112702, they seem to be causingEric Christopher
self host errors on clang-x86-64. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112719 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-01If PrototypeValue is erased in the middle of using the SSAUpdatorDuncan Sands
then the SSAUpdator may access freed memory. Instead, simply pass in the type and name explicitly, which is all that was used anyway. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112699 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-01add a gross hack to work around a problem that Argiris reportedChris Lattner
on llvmdev: SRoA is introducing MMX datatypes like <1 x i64>, which then cause random problems because the X86 backend is producing mmx stuff without inserting proper emms calls. In the short term, force off MMX datatypes. In the long term, the X86 backend should not select generic vector types to MMX registers. This is being worked on, but won't be done in time for 2.8. rdar://8380055 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112696 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-01Revert 112442 and 112440 until the compile time problems introducedDan Gohman
by 112440 are resolved. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112692 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31licm is wasting time hoisting constant foldable operations,Chris Lattner
instead of hoisting them, just fold them away. This occurs in the testcase for PR8041, for example. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112669 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31tidy upChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112643 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31More cleanups of my JumpThreading transforms, including extracting some ↵Owen Anderson
duplicated code into a helper function. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112634 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31Add an RAII helper to make cleanup of the RecursionSet more fool-proof.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112628 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31Only try to clean up the current block if we changed that block already.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112625 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31Refactor my fix for PR5652 to terminate the predecessor lookups after the ↵Owen Anderson
first failure. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112620 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31Fix an infinite loop; merging two functions will create a new function (if theNick Lewycky
two are weak, we make them thunks to a new strong function) so don't iterate through the function list as we're modifying it. Also add back the outermost loop which got removed during the cleanups. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112595 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31Don't perform an extra traversal of the function just to do cleanup. We can ↵Owen Anderson
safely simplify instructions after each block has been processed without worrying about iterator invalidation. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112594 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31Rename ValuePropagation to a more descriptive CorrelatedValuePropagation.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112591 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31Rename file to something more descriptive.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112590 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31More Chris-inspired JumpThreading fixes: use ConstantExpr to correctly ↵Owen Anderson
constant-fold undef, and be more careful with its return value. This actually exposed an infinite recursion bug in ComputeValueKnownInPredecessors which theoretically already existed (in JumpThreading's handling of and/or of i1's), but never manifested before. This patch adds a tracking set to prevent this case. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112589 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31Switch to DenseSet, simplifying much more code. We now have a single iterationNick Lewycky
where we hash, compare and fold, instead of one iteration where we build up the hash buckets and a second one to fold. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112582 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-31Remove r111665, which implemented store-narrowing in InstCombine. Chris ↵Owen Anderson
discovered a miscompilation in it, and it's not easily fixable at the optimizer level. I'll investigate reimplementing it in DAGCombine. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112575 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-30Fix a typo.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112560 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-30Cleanups suggested by Chris.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112553 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-06-23 16:50:04 +0000