aboutsummaryrefslogtreecommitdiff
path: root/lib/Transforms
AgeCommit message (Collapse)Author
2008-04-07Add operator= implementations to SparseBitVector, allowing it to be used in ↵Owen Anderson
GVN. This results in both time and memory savings for GVN. For example, one testcase went from 10.5s to 6s with this patch. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49345 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-07Use Intrinsic::getDeclaration in more places.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49338 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-07The "stacksave is not nounwind problem" no longerDuncan Sands
needs to be fixed here - a previous commit made sure that intrinsics always get the right attributes. So remove no-longer needed code, and while there use Intrinsic::getDeclaration rather than getOrInsertFunction. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49337 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-07Use Intrinsic::getDeclaration to get hold ofDuncan Sands
intrinsics. Fix up the argument type (should be i8*, was an array*). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49336 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-07Make GVN more memory efficient, particularly on code that contains a large ↵Owen Anderson
number of allocations, which GVN can't optimize anyways. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49329 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-07Mark calls to llvm.stacksave, llvm.stackrestore asDale Johannesen
nounwind. When such calls are inlined into something else that is invoked, they were getting changed to invokes, which is badness. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49299 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-06silence a warning when assertions are disabled.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49283 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-06API changes for class Use size reduction, wave 1.Gabor Greif
Specifically, introduction of XXX::Create methods for Users that have a potentially variable number of Uses. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49277 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-02Iterators folloring a SmallVector erased element are invalidated soDavid Greene
don't access cached iterators from after the erased element. Re-apply 49056 with SmallVector support. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49106 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-011. Drop default inline threshold back down to 200.Evan Cheng
2. Do not use # of basic blocks as part of the cost computation since it doesn't really figure into function size. 3. More aggressively inline function with vector code. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49061 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-01Reverting 49056 due to the build being broken.Tanya Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49060 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-01Iterators folloring a SmallVector erased element are invalidated soDavid Greene
don't access cached iterators from after the erased element. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49056 91177308-0d34-0410-b5e6-96231b3b80d8
2008-04-01Revert 49006 for the moment.Dale Johannesen
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49046 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-31Emit exception handling info for functions which areDale Johannesen
not marked nounwind, or for all functions when -enable-eh is set, provided the target supports Dwarf EH. llvm-gcc generates nounwind in the right places; other FEs will need to do so also. Given such a FE, -enable-eh should no longer be needed. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@49006 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-31Don't eliminate bitcast instructions that change the type of a pointerNate Begeman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48971 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-30Fix "Control reaches the end of non-void function" warnings, Chris Lattner
patch by David Chisnall. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48963 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-29change iterator invalidation avoidance to just move the iterator backwardChris Lattner
when something changes, instead of moving forward. This allows us to simplify memset lowering, inserting the memset at the end of the range of stuff we're touching instead of at the start. This, in turn, allows us to make use of the addressing instructions already used in the function instead of inserting our own. For example, we now codegen: %tmp41 = getelementptr [8 x i8]* %ref_idx, i32 0, i32 0 ; <i8*> [#uses=2] call void @llvm.memset.i64( i8* %tmp41, i8 -1, i64 8, i32 1 ) instead of: %tmp20 = getelementptr [8 x i8]* %ref_idx, i32 0, i32 7 ; <i8*> [#uses=1] %ptroffset = getelementptr i8* %tmp20, i64 -7 ; <i8*> [#uses=1] call void @llvm.memset.i64( i8* %ptroffset, i8 -1, i64 8, i32 1 ) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48940 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-29make the common case of a single store (which clearly shouldn't be turnedChris Lattner
into a memset!) faster by avoiding an allocation of an std::list node. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48939 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-29give form-memset a significantly more sane heuristic, enable it by default.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48937 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-28make memset inference significantly more powerful: it can now handle Chris Lattner
memsets that initialize "structs of arrays" and other store sequences that are not sequential. This is still only enabled if you pass -form-memset-from-stores. The flag is not heavily tested and I haven't analyzed the perf regressions when -form-memset-from-stores is passed either, but this causes no make check regressions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48909 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-27PHI->removeIncomingValue may remove PHInode.Devang Patel
Increment iterator in advance. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48890 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-25Handle a special case xor undef, undef -> 0. Technically this should be ↵Evan Cheng
transformed to undef. But this is such a common idiom (misuse) we are going to handle it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48791 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-24Add incoming value from header only if phi node has any use inside the loop.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48738 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-24Increasing the inline limit from (overly conservative) 200 to 300. Given ↵Evan Cheng
each BB costs 20 and each instruction costs 5, 200 means a 4 BB function + 24 instructions (actually less because caller's size also contributes to it). Furthermore, double the limit when more than 10% of the callee instructions are vector instructions. Multimedia kernels tend to love inlining. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48725 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-24Temporarily disabling memset forming optimization. Add an option.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48720 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-24Transform (zext (or (icmp), (icmp))) to (or (zext (cimp), (zext icmp))) if ↵Evan Cheng
at least one of the (zext icmp) can be transformed to eliminate an icmp. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48715 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-23Preserve calling convention during function cloningAnton Korobeynikov
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48708 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-22implement an initial hack at a straight-line store -> memset optimization.Chris Lattner
This fires dozens of times across spec and multisource, but I don't know if it actually speeds stuff up. Hopefully the testers will show something nice :) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48680 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-22implement the logic for memset insertion and store deletion.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48679 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-22This is a partially implemented and currently disabled start of a storeChris Lattner
merging optimization. Nothing to see here, hopefully more later :) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48670 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-21Don't include <map> in Pass.h, which doesn't need it. This requiresDan Gohman
adding <map> to many files that actually do need it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48667 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-21the size of a smallvector shouldn't be part of the interface to these methods.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48662 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-21make gvn marginally faster by reallocating the lastSeenLoad map forChris Lattner
each basic block. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48660 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-21Minor cleanups and shrinkification.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48658 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-21Handle getresult instructions in different basic blocksDan Gohman
from their aggregate operands by moving the getresult instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48657 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-21FunctionExtractorPass has been superceded by GVExtractorPassAndrew Lenharth
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48648 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-21Fix the build for gcc-4.2.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48639 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-21Teach masked value is zero about add and sub, and use MVIZ toChris Lattner
simplify things like (X & 4) >> 1 == 2 --> (X & 4) == 4. since it is obvious that the shift doesn't remove any bits. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48631 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-20Incorporate feedback.Devang Patel
- Fix loop nest. - Use RetVals.size() - Check for null return value. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48605 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-20C and Objective Caml bindings for mem2reg and reg2mem.Gordon Henriksen
Patch by Erick Tryzelaar. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48602 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-20Take the old function's name.Zhou Sheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48588 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-2080 col violation.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48573 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-19Add comment.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48567 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-19Remove dead options.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48556 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-17Update heuritics that estimates cost of call instructions.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48474 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-16C and Objective Caml bindings for several scalar transforms.Gordon Henriksen
Patch originally by Erick Tryzelaar, but has been modified somewhat. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48419 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-14The inst combining of inttoptr into GEP with one index was using the bit size ofBill Wendling
the type instead of the byte size. This was causing troublesome mis-compilations. True to form, this took 2 days to find and is a one-line fix. :-P git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48354 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-13Fix a bug in GVN that Duncan noticed, where we potentially need to insert a Owen Anderson
pointer bitcast when performing return slot optimization. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48343 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-13Update -mem2reg to use succ_iterator instead of iterating across TerminatorInstNick Lewycky
successors. This makes it support nounwind. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48320 91177308-0d34-0410-b5e6-96231b3b80d8
2008-03-12Reimplement the parameter attributes support, phase #1. hilights:Chris Lattner
1. There is now a "PAListPtr" class, which is a smart pointer around the underlying uniqued parameter attribute list object, and manages its refcount. It is now impossible to mess up the refcount. 2. PAListPtr is now the main interface to the underlying object, and the underlying object is now completely opaque. 3. Implementation details like SmallVector and FoldingSet are now no longer part of the interface. 4. You can create a PAListPtr with an arbitrary sequence of ParamAttrsWithIndex's, no need to make a SmallVector of a specific size (you can just use an array or scalar or vector if you wish). 5. All the client code that had to check for a null pointer before dereferencing the pointer is simplified to just access the PAListPtr directly. 6. The interfaces for adding attrs to a list and removing them is a bit simpler. Phase #2 will rename some stuff (e.g. PAListPtr) and do other less invasive changes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@48289 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-07-16 11:42:43 +0000