search for: addressingmodematch

...pare::OptimizeMemoryInst(...) has attracted anyone's attention already. If so, any insight into this problem would be appreciated. If it helps, it's clear that the lion's share of time spent in OptimizeMemoryInst(...) is actually spent matching the address mode of the loads/stores with AddressingModeMatcher::MatchAddr(...). Thanks, Cameron -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20121119/d59a07df/attachment.html>

Hi, I want to improve the way CGP sinks the incoming values of a PHI node towards memory accesses. Improving it means a lot to some of our key benchmarks, and I believe can benefit many general cases as well. CGP's OptimizeMemoryInst function handles PHI nodes by running AddressingModeMatcher on all incoming values to see whether they reach consensus on the addressing mode. It does a reasonably good job in sinking a PHI node if its incoming values are used only once (by the PHI node) and share the same base register and offset. For example, CGP transforms define void @foo(i1 %cond, f...

...pe const*, unsigned int)in libLLVMAnalysis.a(ScalarEvolution.cpp.o) llvm::FastISel::SelectGetElementPtr(llvm::User const*)in libLLVMSelectionDAG.a(FastISel.cpp.o) computeArraySize(llvm::CallInst const*, llvm::TargetData const*, bool)in libLLVMAnalysis.a(MemoryBuiltins.cpp.o) llvm::AddressingModeMatcher::MatchOperationAddr(llvm::User*, unsigned int, unsigned int)in libLLVMTransformUtils.a(AddrModeMatcher.cpp.o) llvm::ComputeMaskedBits(llvm::Value*, llvm::APInt const&, llvm::APInt&, llvm::APInt&, llvm::TargetData const*, unsigned int)in libLLVMAnalysis.a(ValueTracking.cpp.o)...