Chris Lattner b20c0b5092 Enhance the "compare with shift" and "compare with div" ... optimizations to be much more aggressive in the face of exact/nsw/nuw div and shifts. For example, these (which are the same except the first is 'exact' sdiv: define i1 @sdiv_icmp4_exact(i64 %X) nounwind { %A = sdiv exact i64 %X, -5 ; X/-5 == 0 --> x == 0 %B = icmp eq i64 %A, 0 ret i1 %B } define i1 @sdiv_icmp4(i64 %X) nounwind { %A = sdiv i64 %X, -5 ; X/-5 == 0 --> x == 0 %B = icmp eq i64 %A, 0 ret i1 %B } compile down to: define i1 @sdiv_icmp4_exact(i64 %X) nounwind { %1 = icmp eq i64 %X, 0 ret i1 %1 } define i1 @sdiv_icmp4(i64 %X) nounwind { %X.off = add i64 %X, 4 %1 = icmp ult i64 %X.off, 9 ret i1 %1 } This happens when you do something like: (ptr1-ptr2) == 42 where the pointers are pointers to non-unit types. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@125266 91177308-0d34-0410-b5e6-96231b3b80d8		2011-02-10 05:23:05 +00:00
..
CMakeLists.txt	Removed a bunch of unnecessary target_link_libraries.	2010-09-28 22:39:14 +00:00
InstCombine.h	Call SimplifyFDivInst() in InstCombiner::visitFDiv().	2011-01-29 17:50:27 +00:00
InstCombineAddSub.cpp	A bunch of cleanups and simplifications using the new PatternMatch predicates	2011-02-10 05:14:58 +00:00
InstCombineAndOrXor.cpp	more cleanups, notably bitcast isn't used for "signed to unsigned type	2011-02-10 05:17:27 +00:00
InstCombineCalls.cpp	Teach instcombine about the rest of the SSE and SSE2 conversion	2011-01-10 07:19:37 +00:00
InstCombineCasts.cpp	implement an instcombine xform that canonicalizes casts outside of and-with-constant operations.	2011-01-15 06:32:33 +00:00
InstCombineCompares.cpp	Enhance the "compare with shift" and "compare with div"	2011-02-10 05:23:05 +00:00
InstCombineLoadStoreAlloca.cpp	Move getOrEnforceKnownAlignment out of instcombine into Transforms/Utils.	2010-12-25 20:37:57 +00:00
InstCombineMulDivRem.cpp	enhance vmcore to know that udiv's can be exact, and add a trivial	2011-02-06 21:44:57 +00:00
InstCombinePHI.cpp	Teach InstructionSimplify about phi nodes. I chose to have it simply	2010-11-14 13:30:18 +00:00
InstCombineSelect.cpp	Fold select + select where both selects are on the same condition.	2011-01-28 03:28:10 +00:00
InstCombineShifts.cpp	Teach instsimplify some tricks about exact/nuw/nsw shifts.	2011-02-09 17:15:04 +00:00
InstCombineSimplifyDemanded.cpp	teach SimplifyDemandedBits that exact shifts demand the bits they	2011-02-10 05:09:34 +00:00
InstCombineVectorOps.cpp	PR5207: Rename overloaded APInt methods set(), clear(), flip() to	2010-12-01 08:53:58 +00:00
InstCombineWorklist.h	I got tired of VISIBILITY_HIDDEN colliding with the gcc enum. Rename it	2010-05-11 20:16:09 +00:00
InstructionCombining.cpp	Conservatively, clear optional flags, such as nsw, when performing	2011-02-02 02:05:46 +00:00
Makefile	make -fno-rtti the default unless a directory builds with REQUIRES_RTTI.	2010-01-24 20:43:08 +00:00