summaryrefslogtreecommitdiff
path: root/test/Analysis
AgeCommit message (Expand)AuthorFilesLines
2016-05-24[CostModel][X86][XOP] Added XOP costmodel for BITREVERSE Simon Pilgrim1-12/+12
2016-05-19[LAA] Check independence of strided accesses before forward caseMatthew Simpson1-0/+45
2016-05-17[SCEV] Be more aggressive in proving NUWSanjoy Das1-0/+60
2016-05-17[CostModel][X86] Tidied up checksSimon Pilgrim1-125/+45
2016-05-15[CostModel][X86] Added scalar bitreverse testsSimon Pilgrim1-0/+51
2016-05-13[LAA] Include MaxSafeDepDistBytes in the analysis print-outAdam Nemet1-0/+37
2016-05-11[SCEVExpander] Fix a failed cast<> assertionSanjoy Das1-0/+70
2016-05-11[SCEVExpander] Don't break SSA in replaceCongruentIVsSanjoy Das1-0/+41
2016-05-11[SCEV] Be more aggressive around proving no-wrapSanjoy Das1-0/+64
2016-05-11[BasicAA] Compare GEP indices based on value (Fix PR27418)Vedant Kumar1-0/+9
2016-05-10[BasicAA] Guard intrinsics don't write to memorySanjoy Das1-0/+30
2016-05-10[SCEV] Use guards to prove predicatesSanjoy Das1-0/+141
2016-05-09[X86][SSE] Improve cost model for i64 vector comparisons on pre-SSE42 targetsSimon Pilgrim1-8/+8
2016-05-09DivergenceAnalysis: Fix crash with no return blocksMatt Arsenault1-0/+30
2016-05-08[CostModel][X86] Extended comparison instruction cost model tests to include ...Simon Pilgrim1-33/+113
2016-05-07[CostModel][X86] Split BSWAP/BITREVERSE cost tests from CTPOP/CTLZ/CTTZ 'bit ...Simon Pilgrim3-176/+188
2016-05-06[CostModel][X86] Tweak 'SSE2-only' test CPU as it was only disabling SSE41 no...Simon Pilgrim1-1/+1
2016-05-06[CostModel][X86] Added ctlz/cttz undef-zero costmodel testsSimon Pilgrim1-32/+209
2016-05-06[CostModel][X86] Added costmodel tests for vector ctpop/ctlz/cttz/bitreverse/...Simon Pilgrim1-0/+481
2016-05-05[PM] port Branch Frequency Analaysis pass to new PMXinliang David Li10-0/+10
2016-05-05[PM] Port Branch Probability Analysis pass to the new pass manager.Xinliang David Li6-1/+7
2016-05-03[SCEV] Tweak the output format and content of -analyzeSanjoy Das1-10/+10
2016-05-02AMDGPU: llvm.SI.fs.constant is a source of divergenceNicolai Haehnle1-0/+22
2016-05-01[SCEV] When printing via -analysis, dump loop dispositionSanjoy Das1-0/+73
2016-04-29DivergenceAnalysis: Fix crash with unreachable blocksMatt Arsenault1-0/+17
2016-04-27Use DL preferred alignment for alloca in Value::getPointerAlignmentArtur Pilipenko1-1/+7
2016-04-25[SCEV] Improve the run-time checking of the NoWrap predicateSilviu Baranga1-12/+127
2016-04-22Have isKnownNotFullPoison be smarter around control flowSanjoy Das1-0/+74
2016-04-22[X86]: Changing cost for “TRUNCATE v16i32 to v16i8” in SSE4.1 mode.Ashutosh Nema1-1/+1
2016-04-18Port DemandedBits to the new pass manager.Michael Kuperstein1-2/+2
2016-04-18[BPI] Consider deoptimize calls as "unreachable"Sanjoy Das1-0/+20
2016-04-14[DivergenceAnalysis] Treat PHI with incoming undef as constantNicolai Haehnle1-0/+28
2016-04-14[SCEV][LAA] Add tests for SCEV expression transformations performed during LAASilviu Baranga1-0/+21
2016-04-14Revert "Support arbitrary addrspace pointers in masked load/store intrinsics"Adam Nemet1-27/+29
2016-04-13AMDGPU: Remove leftover ShaderType attributes in testsMatt Arsenault4-31/+29
2016-04-12Support arbitrary addrspace pointers in masked load/store intrinsicsArtur Pilipenko1-29/+27
2016-04-11This reverts commit r265913 and r265912Sanjoy Das1-309/+0
2016-04-10[SCEV] See through op.with.overflow intrinsicsSanjoy Das1-0/+309
2016-04-08Re-commit [SCEV] Introduce a guarded backedge taken count and use it in LAA a...Silviu Baranga1-0/+109
2016-04-08Don't IPO over functions that can be de-refinedSanjoy Das1-0/+21
2016-04-06AMDGPU: Add a shader calling conventionNicolai Haehnle1-1/+1
2016-04-06Revert r265535 until we know how we can fix the bots Silviu Baranga1-109/+0
2016-04-06[SCEV] Introduce a guarded backedge taken count and use it in LAA and LVSilviu Baranga1-0/+109
2016-04-05[CFLAA] Fix PR27213; incorrect tagging of args/globalsGeorge Burgess IV2-5/+41
2016-04-04[DependenceAnalysis] Check if result of getConstantPart is nullBrendon Cahoon1-0/+73
2016-03-31[TTI] Let the cost model estimate ctpop costs based on legalityBenjamin Kramer2-8/+19
2016-03-25AMDGPU: Cost model for basic integer operationsMatt Arsenault4-0/+343
2016-03-25AMDGPU: Partially implement getArithmeticInstrCost for FP opsMatt Arsenault4-0/+358
2016-03-25TTI: Report 0 cost for free addrspacecastsMatt Arsenault1-0/+45
2016-03-25TTI: Use 0 for cost of fabs if freeMatt Arsenault1-0/+97