summaryrefslogtreecommitdiffstats
path: root/llvm
Commit message (Collapse)AuthorAgeFilesLines
...
* When lowering an inbounds gep, the intermediate adds can haveChris Lattner2011-02-111-6/+3
| | | | | | | | unsigned overflow (e.g. due to a negative array index), but the scales on array size multiplications are known to not sign wrap. llvm-svn: 125409
* Tidy out asm matcher .inc output.Jim Grosbach2011-02-111-2/+2
| | | | llvm-svn: 125408
* Adds llvm::sys::path::is_separator() to test whether a char is a path separatorZhanyong Wan2011-02-113-12/+32
| | | | | | on the host OS. Reviewed by dgregor. llvm-svn: 125406
* Implement sdiv & udiv for <4 x i16> and <8 x i8> NEON vector types.Nate Begeman2011-02-112-0/+230
| | | | | | This avoids moving each element to the integer register file and calling __divsi3 etc. on it. llvm-svn: 125402
* SimplifySelectOps can only handle selects with a scalar condition. Add a checkNadav Rotem2011-02-111-0/+3
| | | | | | that the condition is not a vector. llvm-svn: 125398
* Fix 9173.Nadav Rotem2011-02-113-0/+67
| | | | | | | Add more folding patterns to constant expressions of vector selects and vector bitcasts. llvm-svn: 125393
* Fix #9190Nadav Rotem2011-02-111-1/+1
| | | | | | | | | | | The bug happens when the DAGCombiner attempts to optimize one of the patterns of the SUB opcode. It tries to create a zero of type v2i64. This type is legal on 32bit machines, but the initializer of this vector (i64) is target dependent. Currently, the initializer attempts to create an i64 zero constant, which fails. Added a flag to tell the DAGCombiner to create a legal zero, if we require that the pass would generate legal types. llvm-svn: 125391
* Test commit.Nadav Rotem2011-02-111-1/+1
| | | | llvm-svn: 125389
* More whitespace cleanup...Jim Grosbach2011-02-112-41/+41
| | | | llvm-svn: 125388
* Remove trailing whitespace.Jim Grosbach2011-02-111-6/+6
| | | | llvm-svn: 125385
* Poison the relational operators ==, !=, <, <=, >=, > on llvm::OptionalDouglas Gregor2011-02-111-0/+54
| | | | | | | objects, since they'll end up using the implicit conversion to "bool" and causing some very "fun" surprises. llvm-svn: 125380
* Make LoopUnswitch preserve ScalarEvolution by just forgetting everything aboutCameron Zwarich2011-02-111-0/+5
| | | | | | | a loop when unswitching it. It only does this in the complex case, because everything should be fine already in the simple case. llvm-svn: 125369
* LoopInstSimplify preserves ScalarEvolution.Cameron Zwarich2011-02-111-0/+1
| | | | llvm-svn: 125368
* fix dumping of METADATA_ATTACHMENT2 names, patch by Peter Housel!Chris Lattner2011-02-111-1/+1
| | | | llvm-svn: 125367
* make ConstantExpr::replaceUsesOfWithOnConstant preserve the inboundsChris Lattner2011-02-111-1/+2
| | | | | | flag. Noticed by Jin Gu Kang! llvm-svn: 125366
* make the constantexpr interfaces for inbounds GEPs follow the same styleChris Lattner2011-02-112-77/+23
| | | | | | as other constantexpr flags, reducing redundancy. llvm-svn: 125365
* Remove std::string version of getNameWithPrefix.Rafael Espindola2011-02-114-21/+10
| | | | llvm-svn: 125363
* Disable this test for now...Daniel Dunbar2011-02-111-0/+3
| | | | llvm-svn: 125361
* Fix buggy fcopysign lowering.Evan Cheng2011-02-112-18/+79
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This define float @foo(float %x, float %y) nounwind readnone { entry: %0 = tail call float @copysignf(float %x, float %y) nounwind readnone ret float %0 } Was compiled to: vmov s0, r1 bic r0, r0, #-2147483648 vmov s1, r0 vcmpe.f32 s0, #0 vmrs apsr_nzcv, fpscr it lt vneglt.f32 s1, s1 vmov r0, s1 bx lr This fails to copy the sign of -0.0f because it's lost during the float to int conversion. Also, it's sub-optimal when the inputs are in GPR registers. Now it uses integer and + or operations when it's profitable. And it's correct! lsrs r1, r1, #31 bfi r0, r1, #31, #1 bx lr rdar://8984306 llvm-svn: 125357
* Remove trailing whitespace.Jim Grosbach2011-02-113-18/+18
| | | | llvm-svn: 125327
* Add a test for the LSR issue exposed by r125254.Cameron Zwarich2011-02-111-0/+32
| | | | llvm-svn: 125325
* Tolerate degenerate phi nodes that can occur in the middle of optimizationNick Lewycky2011-02-102-0/+28
| | | | | | passes. Fixes PR9112. Patch by Jakub Staszak! llvm-svn: 125319
* If we can't avoid running loop-simplify twice for now, at least avoid runningCameron Zwarich2011-02-101-0/+3
| | | | | | iv-users twice. llvm-svn: 125318
* Rename 'loopsimplify' to 'loop-simplify'.Cameron Zwarich2011-02-1021-24/+24
| | | | llvm-svn: 125317
* Add mips o32 tests again with the hope that the buildbot won't complaint againBruno Cardoso Lopes2011-02-101-0/+322
| | | | llvm-svn: 125316
* [AVX] Implement 256-bit vector lowering for SCALAR_TO_VECTOR. ThisDavid Greene2011-02-101-0/+18
| | | | | | | largely completes support for 128-bit fallback lowering for code that is not 256-bit ready. llvm-svn: 125315
* Remove the test to silence the buildbot, will check it in again with a ↵Bruno Cardoso Lopes2011-02-101-321/+0
| | | | | | proper fix soon llvm-svn: 125305
* Clean trailing whitespace.Owen Anderson2011-02-101-20/+20
| | | | llvm-svn: 125304
* Fix a lot of o32 CC issues and add a bunch of tests. Patch by Akira Hatanaka ↵Bruno Cardoso Lopes2011-02-102-47/+392
| | | | | | with some small modifications by me. llvm-svn: 125292
* [AVX] Implement 256-bit vector lowering for EXTRACT_VECTOR_ELT.David Greene2011-02-101-1/+35
| | | | llvm-svn: 125284
* ptx: add passing parameter to kernel functionsChe-Liang Chiou2011-02-109-62/+104
| | | | llvm-svn: 125279
* CMake: LLVM_LIT_TOOLS_DIR is needed only on Win32 hosts to use GnuWin32 tools.NAKAMURA Takumi2011-02-101-2/+3
| | | | | | Unixen and Cygwin do not need it. llvm-svn: 125277
* CMake: LLVM_NO_RTTI must be obsolete now!NAKAMURA Takumi2011-02-101-2/+0
| | | | llvm-svn: 125274
* lit/TestFormats.py: Unittests may be found with suffix .exe also on Cygwin.NAKAMURA Takumi2011-02-101-2/+2
| | | | llvm-svn: 125273
* lit/Util.py: On Cygwin, 'PATHEXT' may exist but it should not be used.NAKAMURA Takumi2011-02-101-1/+5
| | | | llvm-svn: 125272
* implement the first part of PR8882: when lowering an inboundsChris Lattner2011-02-103-10/+33
| | | | | | | | | | | | | | | | | | | | | | | | | | gep to explicit addressing, we know that none of the intermediate computation overflows. This could use review: it seems that the shifts certainly wouldn't overflow, but could the intermediate adds overflow if there is a negative index? Previously the testcase would instcombine to: define i1 @test(i64 %i) { %p1.idx.mask = and i64 %i, 4611686018427387903 %cmp = icmp eq i64 %p1.idx.mask, 1000 ret i1 %cmp } now we get: define i1 @test(i64 %i) { %cmp = icmp eq i64 %i, 1000 ret i1 %cmp } llvm-svn: 125271
* switch the constantexpr, target folder, and IRBuilder interfacesChris Lattner2011-02-105-202/+145
| | | | | | | for NSW/NUW binops to follow the pattern of exact binops. This allows someone to use Builder.CreateAdd(x, y, "tmp", MaybeNUW); llvm-svn: 125270
* Enhance a bunch of transformations in instcombine to start generatingChris Lattner2011-02-1012-160/+196
| | | | | | | | | | | exact/nsw/nuw shifts and have instcombine infer them when it can prove that the relevant properties are true for a given shift without them. Also, a variety of refactoring to use the new patternmatch logic thrown in for good luck. I believe that this takes care of a bunch of related code quality issues attached to PR8862. llvm-svn: 125267
* Enhance the "compare with shift" and "compare with div" Chris Lattner2011-02-103-74/+151
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 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. llvm-svn: 125266
* more cleanups, notably bitcast isn't used for "signed to unsigned type Chris Lattner2011-02-102-46/+28
| | | | | | conversions". :) llvm-svn: 125265
* A bunch of cleanups and simplifications using the new PatternMatch predicatesChris Lattner2011-02-101-176/+132
| | | | | | | | | | and generally tidying things up. Only very trivial functionality changes like now doing (-1 - A) -> (~A) for vectors too. InstCombineAddSub.cpp | 296 +++++++++++++++++++++----------------------------- 1 file changed, 126 insertions(+), 170 deletions(-) llvm-svn: 125264
* teach SimplifyDemandedBits that exact shifts demand the bits they Chris Lattner2011-02-101-3/+23
| | | | | | | are shifting out since they do require them to be zeros. Similarly for NUW/NSW bits of shl llvm-svn: 125263
* After 3-addressifying a two-address instruction, update the register maps; ↵Evan Cheng2011-02-102-14/+31
| | | | | | add a missing check when considering whether it's profitable to commute. rdar://8977508. llvm-svn: 125259
* Revert this in an attempt to bring the builders back.Eric Christopher2011-02-101-4/+4
| | | | llvm-svn: 125257
* Don't return before calling the post-processing function(s).Bill Wendling2011-02-101-1/+0
| | | | llvm-svn: 125256
* Turn this pass ordering:Cameron Zwarich2011-02-101-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | Natural Loop Information Loop Pass Manager Canonicalize natural loops Scalar Evolution Analysis Loop Pass Manager Induction Variable Users Canonicalize natural loops Induction Variable Users Loop Strength Reduction into this: Scalar Evolution Analysis Loop Pass Manager Canonicalize natural loops Induction Variable Users Loop Strength Reduction This fixes <rdar://problem/8869639>. I also filed PR9184 on doing this sort of thing automatically, but it seems easier to just change the ordering of the passes if this is the only case. llvm-svn: 125254
* Do AsmMatcher operand classification per-opcode.Jim Grosbach2011-02-102-89/+29
| | | | | | | | | | When matching operands for a candidate opcode match in the auto-generated AsmMatcher, check each operand against the expected operand match class. Previously, operands were classified independently of the opcode being handled, which led to difficulties when operand match classes were more complicated than simple subclass relationships. llvm-svn: 125245
* Delete unused code for analyzing and splitting around loops.Jakob Stoklund Olesen2011-02-092-389/+2
| | | | | | | Loop splitting is better handled by the more generic global region splitting based on the edge bundle graph. llvm-svn: 125243
* Rip out realpath() support. It's expensive, and often a bad idea, andDouglas Gregor2011-02-092-39/+0
| | | | | | I have another way to achieve the same goal. llvm-svn: 125239
* Simplify using the new leaveIntvBefore()Jakob Stoklund Olesen2011-02-091-13/+2
| | | | llvm-svn: 125238
OpenPOWER on IntegriCloud