Searched refs:DSE (Results 1 – 13 of 13) sorted by relevance
| /external/llvm/test/Transforms/DeadStoreElimination/ |
| D | atomic.ll | 7 ; Note that it turns out essentially every transformation DSE does is legal on
15 ; DSE across unordered store (allowed)
26 ; DSE remove unordered store (allowed)
36 ; DSE unordered store overwriting non-atomic store (allowed)
45 ; DSE no-op unordered atomic store (allowed)
55 ; DSE seq_cst store (be conservative; DSE doesn't have infrastructure
65 ; DSE and seq_cst load (be conservative; DSE doesn't have infrastructure
78 ; DSE across monotonic load (allowed as long as the eliminated store isUnordered)
89 ; DSE across monotonic store (allowed as long as the eliminated store isUnordered)
100 ; DSE across monotonic load (forbidden since the eliminated store is atomic)
[all …]
|
| D | inst-limits.ll | 4 ; If there are two stores to the same location, DSE should be able to remove
15 ; so this store should be optimized away by DSE.
120 ; effect on the working of DSE in any way.
136 ; one too many instruction for the DSE to take place.
|
| D | simple.ll | 128 ; DSE should delete the dead trampoline.
|
| /external/llvm/test/Transforms/EarlyCSE/ |
| D | atomics.ll | 62 ; Can't DSE across a full fence
74 ; Can't DSE across a full fence
86 ; Can't DSE across a full fence
98 ; Can't DSE across a full fence
133 ; Can't DSE a volatile store
143 ; Can't DSE a redundant volatile store
164 ; Can DSE a non-volatile store in favor of a volatile one
175 ; Can DSE a normal store in favor of a unordered one
184 ; Can also DSE a unordered store in favor of a normal one
216 ; Can DSE a unordered store in favor of a unordered one
[all …]
|
| D | fence.ll | 59 ; principal reorder the second store above the fence and then DSE either
60 ; store, but this is beyond the simple last-store DSE which EarlyCSE
74 ; We *could* DSE across this fence, but don't. No other thread can
|
| D | basic.ll | 158 ;; Trivial DSE can't be performed across a readonly call. The call
172 ;; Trivial DSE can be performed across a readnone call.
|
| /external/llvm/lib/Transforms/Scalar/ |
| D | DeadStoreElimination.cpp | 49 struct DSE : public FunctionPass { struct
56 DSE() : FunctionPass(ID), AA(nullptr), MD(nullptr), DT(nullptr) { in DSE() function
101 char DSE::ID = 0;
102 INITIALIZE_PASS_BEGIN(DSE, "dse", "Dead Store Elimination", false, false)
108 INITIALIZE_PASS_END(DSE, "dse", "Dead Store Elimination", false, false) in INITIALIZE_PASS_DEPENDENCY()
110 FunctionPass *llvm::createDeadStoreEliminationPass() { return new DSE(); } in INITIALIZE_PASS_DEPENDENCY()
485 bool DSE::runOnBasicBlock(BasicBlock &BB) { in runOnBasicBlock()
665 bool DSE::MemoryIsNotModifiedBetween(Instruction *FirstI, in MemoryIsNotModifiedBetween()
737 bool DSE::HandleFree(CallInst *F) { in HandleFree()
792 bool DSE::handleEndBlock(BasicBlock &BB) { in handleEndBlock()
[all …]
|
| /external/llvm/docs/ |
| D | Atomics.rst | 241 loads. CSE/DSE and a few other optimizations are allowed, but Monotonic
376 example to do DSE across a single atomic access, but not across a
398 * DSE: Unordered stores can be DSE'ed like normal stores. Monotonic stores can
399 be DSE'ed in some cases, but it's tricky to reason about, and not especially
400 important. It is possible in some case for DSE to operate across a stronger
401 atomic operation, but it is fairly tricky. DSE delegates this reasoning to
|
| D | Lexicon.rst | 85 **DSE**
|
| /external/llvm/test/Analysis/TypeBasedAliasAnalysis/ |
| D | dse.ll | 4 ; DSE should make use of TBAA.
|
| /external/llvm/test/Analysis/BasicAA/ |
| D | phi-aa.ll | 61 ; %0 and %arrayidx5 can alias! It is not safe to DSE the above store.
|
| D | modref.ll | 51 ;; FIXME: DSE isn't zapping this dead store.
|
| /external/llvm/lib/Target/ |
| D | README.txt | 974 DSE should sink partially dead stores to get the store out of the loop.
1688 We should use DSE + llvm.lifetime.end to delete dead vtable pointer updates.
1949 cross-block DSE).
|