aboutsummaryrefslogtreecommitdiff
path: root/lib/Transforms/InstCombine
AgeCommit message (Collapse)Author
2011-07-04PR10267: Don't combine an equality compare with an AND into an inequality ↵Benjamin Kramer
compare when the AND has more than one use. This can pessimize code, inequalities are generally more expensive. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@134379 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-01Generalize @llvm.ctlz, @llvm.cttz, and @llvm.ctpop to work on vectors of ↵Owen Anderson
integers, and fix the one optimization pass that I'm aware of that needs updating for this. At least one current target, ARM NEON, can implement these operations on vectors directly. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@134265 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-23PR10180: Fix a instcombine crash with FP vectors.Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133756 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-20Revamp the "ConstantStruct::get" methods. Previously, these were scatteredChris Lattner
all over the place in different styles and variants. Standardize on two preferred entrypoints: one that takes a StructType and ArrayRef, and one that takes StructType and varargs. In cases where there isn't a struct type convenient, we now add a ConstantStruct::getAnon method (whose name will make more sense after a few more patches land). It would be "really really nice" if the ConstantStruct::get and ConstantVector::get methods didn't make temporary std::vectors. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133412 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-17Revert r133285. Causing odd failures on Dragonegg.Chad Rosier
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133301 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-17Relocate NUW test to cover all binary ops in a dynamic alloca expr.Stuart Hastings
Followup to 132926. rdar://problem/9265821 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133285 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-13Avoid fusing bitcasts with dynamic allocas if the amount-to-allocateStuart Hastings
might overflow. Re-typing the alloca to a larger type (e.g. double) hoists a shift into the alloca, potentially exposing overflow in the expression. rdar://problem/9265821 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132926 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-13InstCombine: Fold A-b == C --> b == A-C if A and C are constants.Benjamin Kramer
The backend already knew this trick. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132915 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-12InstCombine: Shrink ((zext X) & C1) == C2 to fold away the cast if the ↵Benjamin Kramer
"zext" and the "and" have one use. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132897 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-12Simplify code. No functionality changes, name changes aside.Benjamin Kramer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132896 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-01Reapply 132348 with fixes. rdar://problem/6501862Stuart Hastings
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132402 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31Revert to pacify a buildbot. rdar://problem/6501862Stuart Hastings
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132351 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31Followup to 132316; accept arbitrary constants, add with a constant,Stuart Hastings
sub with a non-constant. Fix comments, enlarge test case. rdar://problem/6501862 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132348 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30(1 - X) * (-2) -> (x - 1) * 2, for all positive nonzero powers of 2Stuart Hastings
rdar://problem/6501862 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132316 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-28ConstantFoldInstOperands doesn't like compares, hand it off to instsimplify ↵Benjamin Kramer
instead. Fixes PR10040. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132254 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27InstCombine: Make switch folding with equality compares more aggressive by ↵Benjamin Kramer
trying instsimplify on the arm where we know the compared value. Stuff like "x == y ? y : x&y" now folds into "x&y". git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132185 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27Final step of instcombine debuginfo; switch a couple more places over to ↵Eli Friedman
InsertNewInstWith, and use setDebugLoc for the cases which can't be easily handled by the automated mechanisms. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132167 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-26Renamed llvm.x86.sse42.crc32 intrinsics; crc64 doesn't exist. Chad Rosier
crc32.[8|16|32] have been renamed to .crc32.32.[8|16|32] and crc64.[8|16|32] have been renamed to .crc32.64.[8|64]. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132163 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-25PR9998: ashr exact %x, 31 is not equivalent to sdiv exact %x, -2147483648.Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132097 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-24Make instcombine O(N) instead of O(N^2) in code where the same simplifiable ↵Eli Friedman
constant is used many times. Part of rdar://9471075. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131979 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-23rearrange two transforms, since one subsumes the other. Make the ↵Chris Lattner
shift-exactness xform recurse. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131888 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-23Transform any logical shift of a power of two into an exact/NUW shift whenChris Lattner
in a known-non-zero context. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131887 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-23use the valuetracking isPowerOfTwo function, which is more powerful than ↵Chris Lattner
checking for a constant directly. Thanks to Duncan for pointing this out. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131885 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-22add some random notes.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131862 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-22Carve out a place in instcombine to put transformations which work knowing ↵Chris Lattner
that their result is non-zero. Implement an example optimization (PR9814), which allows us to transform: A / ((1 << B) >>u 2) into: A >>u (B-2) which we compile into: _divu3: ## @divu3 leal -2(%rsi), %ecx shrl %cl, %edi movl %edi, %eax ret instead of: _divu3: ## @divu3 movb %sil, %cl movl $1, %esi shll %cl, %esi shrl $2, %esi movl %edi, %eax xorl %edx, %edx divl %esi, %eax ret git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131860 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-21Revert "InstCombine: Turn mul.with.overflow(X, 2) into the cheaper ↵Benjamin Kramer
add.with.overflow(X, X)" It's better to do this in codegen, mul.with.overflow(X, 2) is more canonical because it has only one use on "X". git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131798 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-21InstCombine: Turn mul.with.overflow(X, 2) into the cheaper ↵Benjamin Kramer
add.with.overflow(X, X) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131789 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-20Revert r131664 and fix it in instcombine instead. rdar://9467055Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131708 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-19Add comment.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131659 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-19Make the demanded bits/elements optimizations preserve debug line information.Eli Friedman
I'm not sure this is quite ideal, but I can't really think of any better way to do it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131616 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-18More instcombine cleanup, towards improving debug line info.Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131604 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-18More instcombine simplifications towards better debug locations.Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131596 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-18More instcombine cleanup aimed towards improving debug line info.Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131559 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-18Switch more inst insertion in instcombine to IRBuilder.Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131547 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-18Switch more inst insertion in instcombine to IRBuilder.Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131544 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-18Switch inst insertion in instcombine transform to IRBuilder.Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131542 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-18Fix inelegant initialization.Stuart Hastings
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131538 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-18Start trying to make InstCombine preserve more debug info. The idea here is ↵Eli Friedman
to set the debug location on the IRBuilder, which will be then right location in most cases. This should magically give many transformations debug locations, and fixing places which are missing a debug location will usually just means changing the code creating it to use the IRBuilder. As an example, the change to InstCombineCalls catches a common case where a call to a bitcast of a function is rewritten. Chris, does this approach look reasonable? git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131516 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-18Use ReplaceInstUsesWith instead of replaceAllUsesWith where appropriate in ↵Eli Friedman
instcombine. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131512 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-17X86 pmovsx/pmovzx ignore the upper half of their inputs.Stuart Hastings
rdar://problem/6945110 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131493 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-14Avoid combining GEPs that might overflow at runtime.Stuart Hastings
rdar://problem/9267970 Patch by Julien Lerouge! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@131339 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-05PR9838: Fix transform introduced in r127064 to not trigger when only one ↵Eli Friedman
side of the icmp is an exact shift. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130954 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-02Remove unused variable.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130705 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-02Move some rem transforms out of instcombine and into instsimplify.Duncan Sands
This automagically provides a transform noticed by my super-optimizer as occurring quite often: "rem x, (select cond, x, 1)" -> 0. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130694 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-30InstCombine: Turn (zext A) udiv (zext B) into (zext (A udiv B)). Same for ↵Benjamin Kramer
urem or constant B. This obviously helps a lot if the division would be turned into a libcall (think i64 udiv on i386), but div is also one of the few remaining instructions on modern CPUs that become more expensive when the bitwidth gets bigger. This also helps register pressure on i386 when dividing chars, divb needs two 8-bit parts of a 16 bit register as input where divl uses two registers. int foo(unsigned char a) { return a/10; } int bar(unsigned char a, unsigned char b) { return a/b; } compiles into (x86_64) _foo: imull $205, %edi, %eax shrl $11, %eax ret _bar: movzbl %dil, %eax divb %sil, %al movzbl %al, %eax ret git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130615 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-30Use SimplifyDemandedBits on div instructions.Benjamin Kramer
This folds away silly stuff like (a&255)/1000 -> 0. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130614 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-29Balance parentheses.Benjamin Kramer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130489 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-29InstCombine: turn (C1 << A) << C2) into (C1 << C2) << A)Benjamin Kramer
Fixes PR9809. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130485 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-28We require threse bits to be zero, too.Benjamin Kramer
This shouldn't happen in practice because the icmp would be a constant. Add a check so we don't miscompile code if something goes wrong. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130446 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-28Fix a comment.Benjamin Kramer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130428 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-08-04 22:14:33 +0000