Only perform DAG combine on FMAs of legal types.
[oota-llvm.git] / lib / Target / X86 / X86ISelLowering.cpp
2012-08-30 Craig TopperOnly perform DAG combine on FMAs of legal types.
2012-08-29 Craig TopperConvert FMA4 patterns to use target specific nodes...
2012-08-28 Michael LiaoAdd comments on the literal value used.
2012-08-28 Michael LiaoExplicitly update the number of nodes to be traversed
2012-08-28 Michael LiaoFix PR12312
2012-08-27 Craig TopperRemove MMX shift intrinsic handling code that also...
2012-08-24 Craig TopperCustom lower FMA intrinsics to target specific nodes...
2012-08-20 Michael Liaofix a case where all operands of BUILD_VECTOR are undefined
2012-08-19 Nadav RotemWhen unsafe math is used, we can use commutative FMAX...
2012-08-18 Nadav RotemReapply r162160 with a fix: Optimize Arith->Trunc-...
2012-08-18 Craig TopperRefactor code a bit to reduce number of calls in the...
2012-08-18 Nadav RotemRevert r162160 because it made a few buildbots fail.
2012-08-18 Nadav RotemThe X86 backend has a number of optimizations for SETCC...
2012-08-17 Craig TopperUse nested switch to select arguments to reduce calls...
2012-08-17 Craig TopperMake ReplaceATOMIC_BINARY_64 a static function. Use...
2012-08-14 Michael Liaominor fix of X86ISD::VSEXT_MOVL dump
2012-08-14 Michael Liaofix PR11334
2012-08-14 Craig TopperFactor duplicate calls to getUNDEF in several functions.
2012-08-14 Craig TopperRe-factor intrinsic lowering to combine common parts...
2012-08-13 Craig TopperTidy up VSETCC lowering code a bit more by adding an...
2012-08-13 Craig TopperRefactor code a bit to share commonalities. No function...
2012-08-13 Craig TopperFix an unused variable warning from r161742.
2012-08-13 Craig TopperRemove the LowerMMXCONCAT_VECTORS function. It could...
2012-08-12 Craig TopperRemove call to setOperationAction for SETCC of v4f32...
2012-08-12 Craig TopperRemove unnecessary call to setOperationAction for SETCC...
2012-08-12 Craig TopperMake replace many calls to getSizeInBits() with is128Bi...
2012-08-12 Craig TopperUse MVT.isXBitVector instead of EVT.isXBitVector when...
2012-08-11 Michael Liaofix PR13577, an issue introduced by r161687
2012-08-11 Craig TopperMove setOperationAction for CONCAT_VECTORS for 256...
2012-08-10 Michael Liaoadd X86-specific DAG optimization to simplify boolean...
2012-08-10 Michael Liaoremove tailing whitespaces and test commit
2012-08-10 Joerg SonnenbergerAdd some missing includes for the build against stdcxx.
2012-08-08 Manman RenX86: enable CSE between CMP and SUB
2012-08-07 Evan ChengX86 cmp lowering is looking past truncate on the condit...
2012-08-06 Craig TopperImplement proper handling for pcmpistri/pcmpestri intri...
2012-08-05 Craig TopperRemove custom inserter for MWAIT. It doesn't do anythin...
2012-08-05 Craig TopperUse a COPY node instead of an explicit MOVA opcode...
2012-08-03 Bob WilsonFall back to selection DAG isel for calls to builtin...
2012-08-01 Chad RosierWhitespace.
2012-08-01 Elena DemikhovskyAdded FMA functionality to X86 target.
2012-07-25 Rafael EspindolaWhen a return struct pointer is passed in registers...
2012-07-23 Sylvestre LedruFix a typo (the the => the)
2012-07-17 Evan ChengBack out r160101 and instead implement a dag combine...
2012-07-17 Evan ChengImplement r160312 as target indepedenet dag combine.
2012-07-17 Evan ChengThis is another case where instcombine demanded bits...
2012-07-16 Evan ChengFor something like
2012-07-15 Nadav RotemTeach getTargetVShiftNode about TargetConstant nodes.
2012-07-14 Nadav RotemAVX: Fix a bug in getTargetVShiftNode. The shift amount...
2012-07-12 Benjamin KramerGive the rdrand instructions a SideEffect flag and...
2012-07-12 Benjamin KramerAdd intrinsics for Ivy Bridge's rdrand instruction.
2012-07-11 Nadav RotemWhen ext-loading and trunc-storing vectors to memory...
2012-07-10 Nadav RotemImprove the loading of load-anyext vectors by allowing...
2012-07-04 Jakob Stoklund OlesenMake X86 call and return instructions non-variadic.
2012-07-04 Jakob Stoklund OlesenEnsure CopyToReg nodes are always glued to the call...
2012-07-01 Elena DemikhovskyOptimization of shuffle node that can fit to the regist...
2012-06-29 Rafael EspindolaIn the initial exec mode we always do a load to find...
2012-06-26 Elena DemikhovskyRemoved unused variable
2012-06-26 Bill WendlingRename to match other X86_64* names.
2012-06-26 Elena DemikhovskyShuffle optimization for AVX/AVX2.
2012-06-25 Eli FriedmanMake some ugly hacks for inline asm operands which...
2012-06-24 Jakob Stoklund Olesen%RCX is not a function live-out in eh.return functions.
2012-06-24 Pete CooperRemove code i'd been testing with but didn't mean to...
2012-06-24 Pete CooperDAG legalisation can now handle illegal fma vector...
2012-06-23 Rafael EspindolaHandle aliases to tls variables in all architectures...
2012-06-20 Craig TopperDon't insert 128-bit UNDEF into 256-bit vectors. Just...
2012-06-19 Rafael EspindolaMove the support for using .init_array from ARM to...
2012-06-09 Craig TopperUse XOP vpcom intrinsics in patterns instead of a targe...
2012-06-09 Craig TopperReplace XOP vpcom intrinsics with fewer intrinsics...
2012-06-08 Manman RenEnable optimization for integer ABS on X86 if Subtarget...
2012-06-07 Manman RenX86: optimize generated code for integer ABS
2012-06-07 Nadav RotemDo not optimize the used bits of the x86 vselect condit...
2012-06-07 Manman RenPR13046: we can't replace usage of SUB with CMP in...
2012-06-07 Manman RenX86: replace SUB with CMP if possible
2012-06-02 Benjamin KramerFix typos found by github.com/lyda/misspell-check
2012-06-01 Hans WennborgImplement the local-dynamic TLS model for x86 (PR3985)
2012-05-31 Jakob Stoklund OlesenAdd support for return value promotion in X86 calling...
2012-05-25 Justin HolewinskiChange interface for TargetLowering::LowerCallTo and...
2012-05-22 Craig TopperFix constant used for pshufb mask when lowering v16i8...
2012-05-21 Craig TopperAllow 256-bit shuffles to still be split even if only...
2012-05-19 Nadav RotemOn Haswell, perfer storing YMM registers using a single...
2012-05-19 Nadav RotemAdd support for additional in-reg vbroadcast patterns
2012-05-18 Craig TopperSimplify code a bit. No functional change intended.
2012-05-18 Craig TopperSimplify handling of v16i8 shuffles and fix a missed...
2012-05-11 Hans WennborgImplement initial-exec TLS model for 32-bit PIC x86
2012-05-10 Nadav RotemFix merge-typo and cleanup
2012-05-10 Nadav RotemAVX2: Add an additional broadcast idiom.
2012-05-10 Nadav RotemGenerate AVX/AVX2 shuffles even when there is a memory...
2012-05-07 Chad RosierFix a regression from r147481. This combine should...
2012-05-07 Manman RenX86: optimization for -(x != 0)
2012-05-06 Craig TopperUse MVT instead of EVT as the argument to all the shuff...
2012-05-06 Craig TopperAdd VPERMQ/VPERMPD to the list of target specific shuff...
2012-05-05 Benjamin KramerAdd a new target hook "predictableSelectIsExpensive".
2012-05-04 Craig TopperFix some loops to match coding standards. No functional...
2012-05-04 Craig TopperFix up some spacing. No functional change.
2012-05-04 Craig TopperSimplify broadcast lowering code. No functional change...
2012-05-04 Craig TopperAllow v16i16 and v32i8 shuffles to be rewritten as...
2012-05-04 Craig TopperSimplify shuffle narrowing code a bit. No functional...
2012-05-03 Craig TopperUse 'unsigned' instead of 'int' in a few places dealing...
2012-05-03 Craig TopperFix 256-bit vpshuflw and vpshufhw immediate encoding...
2012-05-02 Preston GurdFor Intel Atom, use ILP scheduling always, instead...
next