AgeCommit message (Expand)AuthorFilesLines
2015-02-21R600/SI: Use v_madmk_f32madkMatt Arsenault3-5/+233
2015-02-21R600/SI: Try to use v_madak_f32Matt Arsenault3-0/+274
2015-02-21R600/SI: Don't crash when getting immediate operand sizeMatt Arsenault1-0/+7
2015-02-21R600/SI: Fix mad*k definitionsTom Stellard4-2/+40
2015-02-18R600/SI: Remove isel mubuf legalizationTom Stellard3-133/+1
2015-02-18[mips] Avoid redundant sign extension of the result of binary bitwise instruc...Vasileios Kalintiris4-13/+11
2015-02-18X86: Use bitset to manage a bag of bits. NFC.Benjamin Kramer1-7/+3
2015-02-18[mips] [IAS] Fix using .cpsetup with local labels (PR22518).Toma Tabacu4-9/+59
2015-02-18[x86] Tighten the assertions to document that canonicalization hasChandler Carruth1-4/+5
2015-02-18[x86] Switch an if which is trivially true to an assert. NFCChandler Carruth1-14/+14
2015-02-18[x86] Remove some more 'bit' nomenclature from the generic shiftChandler Carruth1-3/+3
2015-02-18[MSan][MIPS] VarArgHelper for MIPS64Mohit K. Bhakkad1-0/+104
2015-02-18[x86] Fold together the two shift lowering strategies. They were doingChandler Carruth1-150/+68
2015-02-18[ARM] Add missing M/R class CPUsBradley Smith2-0/+186
2015-02-18Fixes two issue in SimplifyDemandedBits of sext_in_reg:Michael Kuperstein3-25/+28
2015-02-18[SystemZ] Clean up warningUlrich Weigand1-4/+1
2015-02-18[x86] Refactor the bit shift code the same as I just did the byte shiftChandler Carruth2-24/+36
2015-02-18[SystemZ] Support all TLS access models - CodeGen partUlrich Weigand21-19/+482
2015-02-18Testing commit accessIgor Laevsky1-0/+1
2015-02-18[SystemZ] Support all TLS access models - MC partUlrich Weigand10-21/+344
2015-02-18Reformat.NAKAMURA Takumi4-103/+91
2015-02-18Revert r229622: "[LoopAccesses] Make VectorizerParams global" and others. r22...NAKAMURA Takumi7-570/+169
2015-02-18NFC: Use range-based for loops and more consistent naming.Daniel Jasper1-19/+15
2015-02-18Remove experimental options to control machine block placement.Daniel Jasper2-147/+20
2015-02-18Minor fix after 229495.Elena Demikhovsky1-12/+4
2015-02-18Address post commit review on r229600.Sanjoy Das1-2/+2
2015-02-18AVX-512: Added support for FP instructions with embedded rounding mode.Elena Demikhovsky8-13/+266
2015-02-18[X86] Add another test case for the bug fixed in r229642. With the bug a vpsr...Craig Topper1-0/+15
2015-02-18[x86] Rewrite the byte shift detection to not use boolean variables toChandler Carruth2-22/+53
2015-02-18[X86] Remove AVX512 pslldq/psrldq shift intrinsics. They aren't implemented y...Craig Topper2-22/+0
2015-02-18[X86] Remove AVX2 and SSE2 pslldq and psrldq intrinsics. We can represent the...Craig Topper10-160/+171
2015-02-18Analysis: fix buildbotsSaleem Abdulrasool1-1/+1
2015-02-18[LoopAccesses] Change LAA:getInfo to return a constant referenceAdam Nemet3-12/+16
2015-02-18[LoopAccesses] Add -analyze supportAdam Nemet4-0/+184
2015-02-18[LoopAccesses] Split out LoopAccessReport from VectorizerReportAdam Nemet3-33/+55
2015-02-18[LoopAccesses] Add missing const to APIs in VectorizationReportAdam Nemet3-8/+9
2015-02-18[LoopAccesses] Add canAnalyzeLoopAdam Nemet2-1/+55
2015-02-18[LoopAccesses] Factor out RuntimePointerCheck::needsCheckingAdam Nemet2-9/+22
2015-02-18[LoopAccesses] Change debug messages from LV to LAAAdam Nemet3-43/+46
2015-02-18[LoopAccesses] Modify test to also check symbolic strides with memchecksAdam Nemet1-3/+9
2015-02-18[LoopAccesses] Create the analysis passAdam Nemet5-23/+141
2015-02-18[LoopAccesses] Make blockNeedsPredication staticAdam Nemet3-6/+10
2015-02-18[LoopAccesses] Cache the result of canVectorizeMemoryAdam Nemet3-18/+31
2015-02-18[LoopAccesses] Stash the report from the analysis rather than emitting itAdam Nemet3-9/+22
2015-02-18[LoopAccesses] Make VectorizerParams globalAdam Nemet3-62/+51
2015-02-18[LoopAccesses] Rename LoopAccessAnalysis to LoopAccessInfoAdam Nemet3-35/+34
2015-02-18[LoopAccesses] Make raw_string_ostream local in VectorizationReportAdam Nemet1-6/+5
2015-02-18[InstCombine] Do not insert a GEP instruction before a landingpad instruction.Akira Hatanaka2-4/+48
2015-02-18[BDCE] Don't forget uses of root instructions seen before the instruction itselfHal Finkel2-3/+43
2015-02-18R600/SI: Rename dst encoding field to be consistent with docsMatt Arsenault2-4/+4