Commit Graph

  • 31b4de4e2e Merge with latest master small-data Jack Andersen 2017-01-27 09:29:24 -10:00
  • b6e18354bd Merge branch 'master' into small-data Jack Andersen 2017-01-27 08:16:30 -10:00
  • 1aa15b49aa AMDGPU: Enable FeatureFlatForGlobal on Volcanic Islands master Matt Arsenault 2017-01-27 17:42:26 +00:00
  • 6ea10124d1 [ARM/AArch64] Relocate and update InterleavedAccessPass tests (NFC) Matthew Simpson 2017-01-27 17:33:16 +00:00
  • 1d8c95f0c1 NVPTX: Make NVPTXInferAddressSpaces preserve CFG Matt Arsenault 2017-01-27 17:30:39 +00:00
  • c96f661d59 [CodeGenPrep]No negative cost in the ExtLd promotion Jun Bum Lim 2017-01-27 17:16:37 +00:00
  • 1f3b497b08 [AMDGPU] Turn AMDGPUUnifyMetadata back into module pass Stanislav Mekhanoshin 2017-01-27 16:38:10 +00:00
  • 887da0d8e8 Fix BasicAA incorrect assumption on GEP Mehdi Amini 2017-01-27 16:12:22 +00:00
  • ccbf1f72a6 Avoid using unspecified ordering in MetadataLoader::MetadataLoaderImpl::parseOneMetadata. Ivan Krasin 2017-01-27 15:54:49 +00:00
  • 93c31fb5c9 [mips] Recommit: "N64 static relocation model support" Simon Dardis 2017-01-27 11:36:52 +00:00
  • 2272113e89 [SLP] Refactoring of horizontal reduction analysis, NFC. Alexey Bataev 2017-01-27 10:54:04 +00:00
  • cce6712519 [LICM] When we are recomputing the alias sets for a subloop, we cannot skip sub-subloops. Chandler Carruth 2017-01-27 10:27:32 +00:00
  • 0fa88773d5 [DAGTypeLegalizer] Handle SIGN/ZERO_EXTEND in WidenVecRes_Convert(). Jonas Paulsson 2017-01-27 07:46:26 +00:00
  • 2276fd31f3 [opt-viewer] Introduce global context Adam Nemet 2017-01-27 06:39:09 +00:00
  • 26fff85b5b [opt-viewer] Remove message from the key Adam Nemet 2017-01-27 06:39:08 +00:00
  • b11ef039e0 [opt-viewer] Unique across the different jobs as well Adam Nemet 2017-01-27 06:39:06 +00:00
  • e07991d36b [opt-viewer] Make sorting for the index page deterministic Adam Nemet 2017-01-27 06:39:02 +00:00
  • 2c3a6b52fb [opt-viewer] Include the function in the remark key Adam Nemet 2017-01-27 06:39:01 +00:00
  • 9d0307b0e9 [opt-viewer] Put critical items in parallel Adam Nemet 2017-01-27 06:38:31 +00:00
  • 3142e694e1 Fix unused variable warning. Richard Trieu 2017-01-27 06:06:05 +00:00
  • 611cb2c218 ARM: fix vectorized division on WoA Saleem Abdulrasool 2017-01-27 03:41:53 +00:00
  • fa8de5c26f NewGVN: Add basic dead and redundant store elimination Daniel Berlin 2017-01-27 02:37:11 +00:00
  • b4e1dcdc3f NVPTXCodeGen: Add IPO to libdeps, since r293189. NAKAMURA Takumi 2017-01-27 02:11:10 +00:00
  • e581db178d [APFloat] Reduce some dispatch boilerplates. NFC. Tim Shen 2017-01-27 02:11:07 +00:00
  • a7e6bc8a63 [NVPTX] [InstCombine] Add llvm_unreachable to appease MSVC. Justin Lebar 2017-01-27 02:04:07 +00:00
  • 2420ede9b6 [NVPTX] Fix use-after-stack-free bug in InstCombineCalls. Justin Lebar 2017-01-27 01:49:39 +00:00
  • 6bc6fda8c8 Constant fold switch inst when looking for trivial conditions to unswitch on. Xin Tong 2017-01-27 01:42:20 +00:00
  • 90fe7e78dc [PM] Port LoopLoadElimination to the new pass manager and wire it into the main pipeline. Chandler Carruth 2017-01-27 01:32:26 +00:00
  • d7899341fe [ARM][LegalizerInfo] Specify the type of the opcode. Quentin Colombet 2017-01-27 01:30:46 +00:00
  • a441c0b8b6 [AArch64][LegalizerInfo] Specify the type of the opcode. Quentin Colombet 2017-01-27 01:13:30 +00:00
  • cb244079cc Revert "[AArch64][LegalizerInfo] Specify the type of the initialization list." Quentin Colombet 2017-01-27 01:13:25 +00:00
  • 3e363a8b07 [NVPTX] Upgrade NVVM intrinsics in InstCombineCalls. Justin Lebar 2017-01-27 00:58:58 +00:00
  • 0c722da570 [ValueTracking] Add comment that CannotBeOrderedLessThanZero does the wrong thing for powi. Justin Lebar 2017-01-27 00:58:34 +00:00
  • 0c6a4bdbb7 [LangRef] Make @llvm.sqrt(x) return undef, rather than have UB, for negative x. Justin Lebar 2017-01-27 00:58:03 +00:00
  • f8e4cd6c85 [PM] Flesh out almost all of the late loop passes. Chandler Carruth 2017-01-27 00:50:21 +00:00
  • 49c910dae1 [libFuzzer] simplify the value profiling callback further: don't use (idx MOD prime) on the hot path where it is useless anyway Kostya Serebryany 2017-01-27 00:39:12 +00:00
  • c7fc061f8e [AArch64][LegalizerInfo] Specify the type of the initialization list. Quentin Colombet 2017-01-27 00:39:03 +00:00
  • eff755f3da [libFuzzer] make sure (again) that __builtin_popcountl is compiled into popcnt Kostya Serebryany 2017-01-27 00:20:55 +00:00
  • bcc64b878f [libFuzzer] simplify the value profile code and disable asan/msan on it Kostya Serebryany 2017-01-27 00:09:59 +00:00
  • ba78b9089a NFC: Rename PDB_ReaderType::Raw to Native for consistency with the NativeSession rename. Adrian McCarthy 2017-01-27 00:01:55 +00:00
  • 566dac5487 Switch the default for building GlobalISel. Quentin Colombet 2017-01-26 23:53:31 +00:00
  • eca8aa608d CMake is funky on detecting Intel 17 as GCC compatible. Yichao Yu 2017-01-26 23:50:18 +00:00
  • 6159a002c7 [ARM] Fix some Clang-tidy modernize and Include What You Use warnings; other minor fixes (NFC). Eugene Zelenko 2017-01-26 23:40:06 +00:00
  • ac8bbffa16 GlobalISel: support debug intrinsics. Tim Northover 2017-01-26 23:39:14 +00:00
  • 78fa56ce60 Revert a couple of InstCombine/Guard checkins Sanjoy Das 2017-01-26 23:38:11 +00:00
  • e6f10d3481 Add intrinsics for constrained floating point operations Andrew Kaylor 2017-01-26 23:27:59 +00:00
  • e1c01ccd11 [PM] Enable the main loop pass pipelines with everything but loop-unswitch in the main pipelines for the new PM. Chandler Carruth 2017-01-26 23:21:17 +00:00
  • bf34add9c8 [obj2yaml] Produce correct output for invalid relocations. Davide Italiano 2017-01-26 23:12:53 +00:00
  • f7b66eb71e [Hexagon] Require IPO library in Hexagon build Krzysztof Parzyszek 2017-01-26 23:03:22 +00:00
  • 8dc127c160 NewGVN: Fix bug exposed by PR31761 Daniel Berlin 2017-01-26 22:21:48 +00:00
  • 4eb218a6fd [InstCombine] fold (X >>u C) << C --> X & (-1 << C) Sanjay Patel 2017-01-26 22:08:10 +00:00
  • a28da41a4a [GlobalISel] Remove duplicate function using variadic templates. NFC. Ahmed Bougacha 2017-01-26 22:07:37 +00:00
  • e264bcdc06 [Hexagon] Add Hexagon-specific loop idiom recognition pass Krzysztof Parzyszek 2017-01-26 21:41:10 +00:00
  • a3fb26f74c NewGVN: Add algorithm overview Daniel Berlin 2017-01-26 21:39:49 +00:00
  • 6a295f2666 [InstCombine] use m_APInt to allow (X << C) >>u C --> X & (-1 >>u C) with splat vectors Sanjay Patel 2017-01-26 20:52:27 +00:00
  • 8f4444960f [Doc][LangRef] Fix typo-ish error in description of Masked Gather Zvi Rackover 2017-01-26 20:29:15 +00:00
  • 76aa95eea9 [InstCombine] add tests for shift-shift folds; NFC Sanjay Patel 2017-01-26 20:10:55 +00:00
  • 04acf641c1 [AArch64] Refine Kryo Machine Model Balaram Makam 2017-01-26 20:10:41 +00:00
  • 68ce9328c9 [IfConversion] Use reverse_iterator to simplify. NFC Kyle Butt 2017-01-26 20:02:47 +00:00
  • eceaa2ad0b [PPC] cleanup of mayLoad/mayStore flags and memory operands. Sean Fertile 2017-01-26 18:59:15 +00:00
  • b0b549acf5 NewGVN: Fix output of pr31578 testcase now that we mark unreachable blocks as unreachable Daniel Berlin 2017-01-26 18:49:03 +00:00
  • c548ce34ef NewGVN: Make unreachable blocks be marked with unreachable Daniel Berlin 2017-01-26 18:30:29 +00:00
  • be4948cead Replace addEarlyAsPossiblePasses callback with adjustPassManager Stanislav Mekhanoshin 2017-01-26 16:49:08 +00:00
  • d9031ef908 Revert "In visitSTORE, always use FindBetterChain, rather than only when UseAA is enabled." Nirav Dave 2017-01-26 16:46:13 +00:00
  • eab2fdc64a [XRay][Arm32] Reduce the portion of the stub and implement more staging for tail calls - in LLVM Serge Rogatch 2017-01-26 16:17:03 +00:00
  • dbb7a65598 In visitSTORE, always use FindBetterChain, rather than only when UseAA is enabled. Nirav Dave 2017-01-26 16:02:24 +00:00
  • dbe38eb993 Use shouldAssumeDSOLocal in classifyGlobalReference. Rafael Espindola 2017-01-26 15:02:31 +00:00
  • 10559c6633 [X86][SSE] Add support for combining ANDNP byte masks with target shuffles Simon Pilgrim 2017-01-26 14:31:12 +00:00
  • 783633e29e [SCEV] Introduce add operation inlining limit Daniil Fukalov 2017-01-26 13:33:17 +00:00
  • b81877488f [X86][SSE] Pull out target shuffle resolve code into helper. NFCI. Simon Pilgrim 2017-01-26 13:06:02 +00:00
  • c2598d4df6 Remove a '#if 0' that wasn't intended for commit in r293173. Daniel Sanders 2017-01-26 12:10:43 +00:00
  • dbafc59fbf Attempt to fix windows buildbots after r293172. Daniel Sanders 2017-01-26 11:23:49 +00:00
  • 2b464c4dc4 [globalisel] Re-factor ISel matchers into a hierarchy. NFC Daniel Sanders 2017-01-26 11:10:14 +00:00
  • 9016ce4442 [AMDGPU] Fix typo in GCNSchedStrategy Valery Pykhtin 2017-01-26 10:51:47 +00:00
  • 5db8081bfd Revert "[mips] N64 static relocation model support" Simon Dardis 2017-01-26 10:46:07 +00:00
  • faae5327de [LV] Fix an issue where forming LCSSA in the place that we did would change the set of uniform instructions in the loop causing an assert failure. Chandler Carruth 2017-01-26 10:41:09 +00:00
  • 40f771f073 [mips] N64 static relocation model support Simon Dardis 2017-01-26 10:19:02 +00:00
  • 4cead7216b [ARM] GlobalISel: Load i1, i8 and i16 args from stack Diana Picus 2017-01-26 09:20:47 +00:00
  • 69c8d681d3 [SLP] Add one more reduction operation for extra argument test to make it vectorizable. Alexey Bataev 2017-01-26 09:18:41 +00:00
  • c5374d2ba5 [PM] Use PoisoningVH correctly when merely deleting entries in a map with it. Chandler Carruth 2017-01-26 08:31:54 +00:00
  • 82579da889 Chapter3/KaleidoscopeJIT.h: Fix a warning. [-Wunused-lambda-capture] NAKAMURA Takumi 2017-01-26 08:31:14 +00:00
  • ac9b2d97ef [TargetTransformInfo] Add override keywords to supporess -Winconsistent-missing-override. Craig Topper 2017-01-26 08:04:27 +00:00
  • facc9962a6 [AVX-512] Move the combine that runs combineBitcastForMaskedOp to the last DAG combine phase where I had originally meant to put it. Craig Topper 2017-01-26 07:17:58 +00:00
  • 3cd2419e8b [X86] When bitcasting INSERT_SUBVECTOR/EXTRACT_SUBVECTOR to match masked operations, use the correct type for the immediate operand. Craig Topper 2017-01-26 07:17:53 +00:00
  • 7cb6abb7cb [TargetTransformInfo] Refactor and improve getScalarizationOverhead() Jonas Paulsson 2017-01-26 07:03:25 +00:00
  • add597c306 [SLP] Fixed test for extra arguments in horizontal reductions. Alexey Bataev 2017-01-26 06:19:52 +00:00
  • c3374e91f1 [DAGCombiner] Fold extract_subvector of undef to undef. Fold away inserting undef subvectors. Craig Topper 2017-01-26 05:38:46 +00:00
  • fafe30250c [X86] Add demanded elts support for the inputs to pclmul intrinsic Craig Topper 2017-01-26 05:17:13 +00:00
  • c41f271a87 Revert test commit Taewook Oh 2017-01-26 04:34:25 +00:00
  • c4a35ed88f test commit Taewook Oh 2017-01-26 04:32:40 +00:00
  • 33e0aba0c0 [OptDiag] Predicates to check the same type of IR and MIR opt remarks Adam Nemet 2017-01-26 04:03:18 +00:00
  • 988a4218f4 gold-plugin: Fix test case. Peter Collingbourne 2017-01-26 02:15:08 +00:00
  • 6d7e8f10e2 [PM] Simplify the new PM interface to the loop unroller and expose two factory functions for the two modes the loop unroller is actually used in in-tree: simplified full-unrolling and the entire thing including partial unrolling. Chandler Carruth 2017-01-26 02:13:50 +00:00
  • a50ea04aad [Loops] Restructure the LoopInfo verify function so that it more directly walks the current loop structure verifying that a matching structure can be found in a freshly computed version. Chandler Carruth 2017-01-26 02:07:20 +00:00
  • 289792dbca gold-plugin: Simplify naming of object files created with save-temps or obj-path. Peter Collingbourne 2017-01-26 02:07:05 +00:00
  • f544b74edf Fix --Wunused-function. Rui Ueyama 2017-01-26 02:03:58 +00:00
  • 6c2c7c89da [libFuzzer] remove a bit of stale code Kostya Serebryany 2017-01-26 01:45:54 +00:00
  • c43e0682f2 [libFuzzer] further simplify __sanitizer_cov_trace_pc_guard Kostya Serebryany 2017-01-26 01:34:58 +00:00
  • baa08a9804 AMDGPU: Fold fneg into round instructions Matt Arsenault 2017-01-26 01:25:36 +00:00
  • 93716868f3 [ImplicitNullChecks] Add a test demonstrating a case we don't get today Sanjoy Das 2017-01-26 01:07:33 +00:00