Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Move createReversePostOrderFunctionAttrsPass right after the inliner is done | Mehdi Amini | 2016-05-02 | 2 | -8/+11 |
| | | | | | | | | | | This is where it was originally, until LoopVersioningLICM was inserted before in r259986, I don't believe it was on purpose. Differential Revision: http://reviews.llvm.org/D19809 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 268252 | ||||
* | [LLE] Fix typo from r263058 | Adam Nemet | 2016-05-02 | 1 | -1/+1 |
| | | | | | | | | This was meant to check unit stride for both the load and the store. Thanks to Roman Shirokiy for noticing this. llvm-svn: 268251 | ||||
* | Add llvm-pdbdump to the tool substitutions list in lit. NFC. | Pete Cooper | 2016-05-02 | 1 | -0/+1 |
| | | | | | | | | | This adds llvm-pdbdump to the list of tools which get printed with the full path in verbose mode. This makes it easier to take the whole run line from verbose output and run it again without prepending with the builds bin directory. llvm-svn: 268250 | ||||
* | Fixed signed/unsigned comparison warning. | Simon Pilgrim | 2016-05-02 | 1 | -1/+1 |
| | | | | llvm-svn: 268249 | ||||
* | Remove extra whitespace. NFC. | Chad Rosier | 2016-05-02 | 1 | -4/+4 |
| | | | | llvm-svn: 268248 | ||||
* | AMDGPU/SI: Use hazard recognizer to detect DPP hazards | Tom Stellard | 2016-05-02 | 4 | -57/+33 |
| | | | | | | | | | | Reviewers: arsenm Subscribers: arsenm, llvm-commits Differential Revision: http://reviews.llvm.org/D18603 llvm-svn: 268247 | ||||
* | remove blank lines | Sanjay Patel | 2016-05-02 | 1 | -8/+1 |
| | | | | llvm-svn: 268246 | ||||
* | [InstCombine] regenerate checks | Sanjay Patel | 2016-05-02 | 1 | -342/+493 |
| | | | | llvm-svn: 268245 | ||||
* | [InstCombine] regenerate checks | Sanjay Patel | 2016-05-02 | 1 | -81/+73 |
| | | | | llvm-svn: 268244 | ||||
* | [InstCombine] regenerate checks | Sanjay Patel | 2016-05-02 | 1 | -69/+103 |
| | | | | llvm-svn: 268242 | ||||
* | [InstCombine] regenerate checks | Sanjay Patel | 2016-05-02 | 1 | -131/+183 |
| | | | | llvm-svn: 268241 | ||||
* | [InstCombine] regenerate checks | Sanjay Patel | 2016-05-02 | 1 | -33/+43 |
| | | | | llvm-svn: 268239 | ||||
* | [X86][SSE] Dropped X86ISD::FGETSIGNx86 and use MOVMSK instead for FGETSIGN ↵ | Simon Pilgrim | 2016-05-02 | 4 | -37/+12 |
| | | | | | | | | lowering movmsk.ll tests are unchanged. llvm-svn: 268237 | ||||
* | Cleanup comments. NFC. | Chad Rosier | 2016-05-02 | 2 | -3/+4 |
| | | | | llvm-svn: 268236 | ||||
* | Cleanup comments. NFC. | Chad Rosier | 2016-05-02 | 1 | -4/+3 |
| | | | | llvm-svn: 268235 | ||||
* | Silence unused variable warnings; NFC. | Aaron Ballman | 2016-05-02 | 1 | -9/+4 |
| | | | | llvm-svn: 268234 | ||||
* | Cleanup comments. NFC. | Chad Rosier | 2016-05-02 | 1 | -7/+9 |
| | | | | llvm-svn: 268233 | ||||
* | [InstCombine] regenerate checks | Sanjay Patel | 2016-05-02 | 1 | -500/+674 |
| | | | | llvm-svn: 268232 | ||||
* | Enable the X86 call frame optimization for the 64-bit targets that allow it. | David L Kreitzer | 2016-05-02 | 3 | -16/+229 |
| | | | | | | | | Fixes PR27241. Differential Revision: http://reviews.llvm.org/D19688 llvm-svn: 268227 | ||||
* | Expose a getFullName for thin archive members. | Rafael Espindola | 2016-05-02 | 2 | -10/+19 |
| | | | | | | It will be used in lld. llvm-svn: 268226 | ||||
* | [SystemZ] Temporarily disable codegen test int-add-12.ll. | Jonas Paulsson | 2016-05-02 | 1 | -1/+1 |
| | | | | | | This checks for AGSI transformation, which is temporarily disabled. llvm-svn: 268219 | ||||
* | [SystemZ] Fix in restoreCalleeSavedRegisters() | Jonas Paulsson | 2016-05-02 | 1 | -1/+2 |
| | | | | | | | | Only add operands for GRs to the LMG. Reviewed by Ulrich Weigand. llvm-svn: 268216 | ||||
* | [SystemZ] Mark CC defs as dead whenever possible. | Jonas Paulsson | 2016-05-02 | 3 | -5/+25 |
| | | | | | | | | | | | | | | Marking implicit CC defs as dead everywhere except when CC is actually defined and used explicitly, is important since the post-ra scheduler will otherwise insert edges between instructions unnecessarily. Also temporarily disable LA(Y)-> AGSI optimization in foldMemoryOperandImpl(), since this inroduces a def of the CC reg, which is illegal unless it is known to be dead. Reviewed by Ulrich Weigand. llvm-svn: 268215 | ||||
* | [X86] Fix a bug in LOCK arithmetic operation pattern matching where the ↵ | Craig Topper | 2016-05-02 | 1 | -1/+1 |
| | | | | | | | | wrong immediate predicate check was being used for 64-bit instructions with 8-bit immediates. This didn't cause a bug because the order of the patterns ensured that the 64-bit instructions with 32-bit immediates were selected first. llvm-svn: 268212 | ||||
* | Fix grammar and correct comment - the debug information wasn't incorrect, ↵ | Eric Christopher | 2016-05-02 | 1 | -2/+2 |
| | | | | | | rather suboptimal. llvm-svn: 268211 | ||||
* | [llvm-readobj] Dump hash as part of -version-info. | Davide Italiano | 2016-05-02 | 2 | -0/+4 |
| | | | | llvm-svn: 268210 | ||||
* | [CodeGen] Add OPC_MoveChild0-OPC_MoveChild7 opcodes to isel matching tables ↵ | Craig Topper | 2016-05-02 | 3 | -3/+24 |
| | | | | | | to optimize table size. Shaves about 12K off the X86 matcher table. llvm-svn: 268209 | ||||
* | [GlobalDCE] Modernize. Use FileCheck instead of grep. | Davide Italiano | 2016-05-01 | 6 | -7/+15 |
| | | | | llvm-svn: 268207 | ||||
* | [InstCombine][SSE] Added support to VPERMD/VPERMPS to shuffle combine to ↵ | Simon Pilgrim | 2016-05-01 | 2 | -13/+18 |
| | | | | | | accept UNDEF elements. llvm-svn: 268206 | ||||
* | Dropped FIXME comment | Simon Pilgrim | 2016-05-01 | 1 | -1/+1 |
| | | | | llvm-svn: 268205 | ||||
* | [InstCombine][SSE] Added support to VPERMILVAR to shuffle combine to accept ↵ | Simon Pilgrim | 2016-05-01 | 2 | -28/+35 |
| | | | | | | UNDEF elements. llvm-svn: 268204 | ||||
* | [InstCombine][AVX] Fixed PERMILVAR identity tests and added additional ↵ | Simon Pilgrim | 2016-05-01 | 1 | -13/+47 |
| | | | | | | decode tests llvm-svn: 268203 | ||||
* | [InstCombine][SSE] Added support to PSHUFB to shuffle combine to accept ↵ | Simon Pilgrim | 2016-05-01 | 2 | -22/+21 |
| | | | | | | UNDEF elements. llvm-svn: 268202 | ||||
* | [InstCombine][SSE] Regenerate MOVSX/MOVZX tests | Simon Pilgrim | 2016-05-01 | 2 | -114/+116 |
| | | | | llvm-svn: 268201 | ||||
* | [AVX512] VPACKUSWB/VPACKSSWB should not be encoded with EVEX.W=1. While ↵ | Craig Topper | 2016-05-01 | 4 | -214/+214 |
| | | | | | | there fix the execution domain for VPACKSSDW/VPACKUSDW. llvm-svn: 268200 | ||||
* | [InstCombine][AVX2] Combine VPERMD/VPERMPS intrinsics with constant masks to ↵ | Simon Pilgrim | 2016-05-01 | 2 | -15/+50 |
| | | | | | | shufflevector. llvm-svn: 268199 | ||||
* | Fixed MSVC 'not all control paths return a value' warning | Simon Pilgrim | 2016-05-01 | 1 | -0/+1 |
| | | | | llvm-svn: 268198 | ||||
* | Document the LLVM_ENABLE_EXPENSIVE_CHECKS cmake option introduced in r268050 | Simon Pilgrim | 2016-05-01 | 1 | -0/+3 |
| | | | | llvm-svn: 268197 | ||||
* | getelementptr instruction, support index vector of EVT. | Igor Breger | 2016-05-01 | 2 | -1/+11 |
| | | | | | | Differential Revision: http://reviews.llvm.org/D19775 llvm-svn: 268195 | ||||
* | Change AVX512 braodcastsd/ss patterns interaction with spilling . New ↵ | Igor Breger | 2016-05-01 | 6 | -110/+329 |
| | | | | | | | | implementation take a scalar register and generate a vector without COPY_TO_REGCLASS (turn it into a VR128 register ) .The issue is that during register allocation we may spill a scalar value using 128-bit loads and stores, wasting cache bandwidth. Differential Revision: http://reviews.llvm.org/D19579 llvm-svn: 268190 | ||||
* | [AVX512] Prefer AVX512 VPACK instructions over AVX/AVX2 instructions when ↵ | Craig Topper | 2016-05-01 | 2 | -11/+11 |
| | | | | | | VLX and BWI are supported. llvm-svn: 268189 | ||||
* | [AVX512] Add HasVLX to the 128/256-bit versions of VPACKSSDW/USDW/SSWB/USWB ↵ | Craig Topper | 2016-05-01 | 1 | -13/+14 |
| | | | | | | and VPMADDUBSW/VPMADDWD. llvm-svn: 268188 | ||||
* | [AVX512] Make sure 128/256-bit DQI versions of VAND/VANDN/VOR/VXOR are also ↵ | Craig Topper | 2016-05-01 | 1 | -16/+16 |
| | | | | | | marked as requiring VLX. llvm-svn: 268186 | ||||
* | [X86] Add an AddedComplexity to another pattern to put it near similar in ↵ | Craig Topper | 2016-05-01 | 1 | -2/+1 |
| | | | | | | the output file. llvm-svn: 268184 | ||||
* | [X86] Remove a seemlingly unused pattern. The same pattern appears elsewhere ↵ | Craig Topper | 2016-05-01 | 1 | -2/+0 |
| | | | | | | with an AddedComplexity that made this unreachable. llvm-svn: 268183 | ||||
* | [X86] Add AddedComplexity to keep some similar patterns near each other in ↵ | Craig Topper | 2016-05-01 | 1 | -0/+1 |
| | | | | | | the output file. llvm-svn: 268181 | ||||
* | [X86] Remove some redundant selection patterns. | Craig Topper | 2016-05-01 | 2 | -11/+0 |
| | | | | llvm-svn: 268180 | ||||
* | [AVX512] Replace vector_extract with extractelt in some patterns. They mean ↵ | Craig Topper | 2016-05-01 | 1 | -5/+5 |
| | | | | | | the same thing but vector_extract is deprecated. NFC llvm-svn: 268179 | ||||
* | [SCEV] When printing via -analysis, dump loop disposition | Sanjoy Das | 2016-05-01 | 2 | -0/+98 |
| | | | | | | | | | | | There are currently some bugs in tree around SCEV caching an incorrect loop disposition. Printing out loop dispositions will let us write whitebox tests as those are fixed. The dispositions are printed as a list in "inside out" order, i.e. innermost loop first. llvm-svn: 268177 | ||||
* | Properly name LLVMSetIsInBounds's argument. NFC | Amaury Sechet | 2016-05-01 | 2 | -3/+3 |
| | | | | llvm-svn: 268176 |