projects / oota-llvm.git / search
? search:
summary | shortlog | log | commit | commitdiff | tree
first ⋅ prev ⋅ next
[GlobalsAA] Turn GlobalsAA on again by default
2015-10-13 James Molloy[GlobalsAA] Turn GlobalsAA on again by default
 commit | commitdiff | tree
2015-10-13 James Molloy[GlobalsAA] Don't assume anything about functions that...
commit | commitdiff | tree
2015-10-12 James Molloy[ARM] Mark Swift MISched model as incomplete
 commit | commitdiff | tree
2015-10-12 James Molloy[LoopVectorize] Shrink integer operations into the...
commit | commitdiff | tree
2015-10-12 James Molloy[MISched] Python script to check coverage of misched...
commit | commitdiff | tree
2015-10-08 James MolloyCompute demanded bits for icmp instructions
 commit | commitdiff | tree
2015-10-08 James MolloyTreat Mul just like Add and Subtract
 commit | commitdiff | tree
2015-10-08 James MolloyMake demanded bits lazy
 commit | commitdiff | tree
2015-10-07 James MolloyRevert "This patch builds on top of D13378 to handle...
commit | commitdiff | tree
2015-09-29 James Molloy[ValueTracking] Teach isKnownNonZero about monotonically...
commit | commitdiff | tree
2015-09-25 James Molloy[GlobalsAA] Teach GlobalsAA about nocapture
 commit | commitdiff | tree
2015-09-24 James Molloy[ValueTracking] Teach isKnownNonZero a new trick
 commit | commitdiff | tree
2015-09-21 James Molloy[LoopUtils,LV] Propagate fast-math flags on generated...
commit | commitdiff | tree
2015-09-21 James MolloyRevert "[ARM] Handle +t2dsp feature as an ArchExtKind...
commit | commitdiff | tree
2015-09-15 James Molloy[GlobalsAA] Disable globals-aa by default
 commit | commitdiff | tree
2015-09-10 James MolloyEnable GlobalsAA by default
 commit | commitdiff | tree
2015-09-10 James MolloyAdd GlobalsAA as preserved to a bunch of transforms
 commit | commitdiff | tree
2015-09-10 James Molloy[ARM] Do not use vtrn for vectorshuffle if the order...
commit | commitdiff | tree
2015-09-09 James MolloyRename ExitCount to BackedgeTakenCount, because that...
commit | commitdiff | tree
2015-09-09 James MolloyDelay predication of stores until near the end of vector...
commit | commitdiff | tree
2015-09-02 James Molloy[ValueTracking] Minor comment change in test
 commit | commitdiff | tree
2015-09-02 James Molloy[ValueTracking] Look through casts when both operands...
commit | commitdiff | tree
2015-09-02 James Molloy[LV] Don't bail to MiddleBlock if a runtime check fails...
commit | commitdiff | tree
2015-09-02 James Molloy[LV] Move some code around slightly to make the intent...
commit | commitdiff | tree
2015-09-02 James Molloy[LV] Cleanup: Sink an IRBuilder closer to its uses.
 commit | commitdiff | tree
2015-09-02 James Molloy[LV] Refactor all runtime check emissions into helper...
commit | commitdiff | tree
2015-09-02 James Molloy[LV] Pull creation of trip counts into a helper function.
 commit | commitdiff | tree
2015-09-02 James Molloy[LV] Factor the creation of the loop induction variable...
commit | commitdiff | tree
2015-09-02 James Molloy[LV] Never widen an induction variable.
 commit | commitdiff | tree
2015-09-02 James Molloy[LV] Switch to using canonical induction variables.
 commit | commitdiff | tree
2015-08-29 James Molloy[ARM] Fix up buildbots after r246360
 commit | commitdiff | tree
2015-08-29 James Molloy[ARM] Hoist fabs/fneg above a conversion to float.
 commit | commitdiff | tree
2015-08-28 James Molloy[LoopUtils] Move a private constructor nearer the other...
commit | commitdiff | tree
2015-08-27 James Molloy[LoopVectorize] Extract InductionInfo into a helper...
commit | commitdiff | tree
2015-08-20 James Molloy[ARM] Don't try and custom lower a vNi64 SETCC.
 commit | commitdiff | tree
2015-08-17 James Molloy[ARM] Fix crash when targetting CPU without NEON
 commit | commitdiff | tree
2015-08-17 James MolloyRemove hand-rolled matching for fmin and fmax.
 commit | commitdiff | tree
2015-08-17 James MolloyRip out hand-rolled matching code for VMIN, VMAX, VMINNM...
commit | commitdiff | tree
2015-08-17 James MolloyGenerate FMINNAN/FMINNUM/FMAXNAN/FMAXNUM from SDAGBuilder.
 commit | commitdiff | tree
2015-08-14 James MolloySeparate out BDCE's analysis into a separate DemandedBits...
commit | commitdiff | tree
2015-08-14 James Molloy[AArch64] FMINNAN/FMAXNAN on f16 is not legal.
 commit | commitdiff | tree
2015-08-13 James Molloy[ARM] FMINNAN/FMAXNAN of f64 are not legal.
 commit | commitdiff | tree
2015-08-13 James Molloy[ARM] Allow vmin/vmax of scalars to be emitted without...
commit | commitdiff | tree
2015-08-13 James Molloy[ARM] Rejig vmax tests a bit
 commit | commitdiff | tree
2015-08-13 James Molloy[AArch64] Small rejig of fmax tests, NFCI.
 commit | commitdiff | tree
2015-08-12 James Molloy[ValueTracking] Tweak a comment slightly
 commit | commitdiff | tree
2015-08-11 James Molloy[AArch64] Match fminnum/fmaxnum for vector fminnm/fmaxnm...
commit | commitdiff | tree
2015-08-11 James Molloy[AArch64] Replace the custom AArch64ISD::FMIN/MAX nodes...
commit | commitdiff | tree
2015-08-11 James Molloy[ARM] Match fminnan/fmaxnan for vector vmin/vmax instead...
commit | commitdiff | tree
2015-08-11 James Molloy[ARM] Match fminnum/fmaxnum for vector vminnm/vmaxnm...
commit | commitdiff | tree
2015-08-11 James Molloy[ARM] Replace ARMISD::VMINNM/VMAXNM with ISD::FMINNUM...
commit | commitdiff | tree
2015-08-11 James Molloy[ARM] Replace ARMISD::FMIN/FMAX with the shiny new...
commit | commitdiff | tree
2015-08-11 James MolloyAdd new ISD nodes: ISD::FMINNAN and ISD::FMAXNAN
 commit | commitdiff | tree
2015-08-11 James MolloyAdd support for floating-point minnum and maxnum
 commit | commitdiff | tree
2015-08-03 James MolloyBe less conservative about forming IT blocks.
 commit | commitdiff | tree
2015-07-17 James Molloy[ARM] Use [SU]ABSDIFF nodes instead of intrinsics for...
commit | commitdiff | tree
2015-07-17 James Molloy[AArch64] Use [SU]ABSDIFF nodes instead of intrinsics...
commit | commitdiff | tree
2015-07-16 James Molloy[Codegen] Add intrinsics 'absdiff' and corresponding...
commit | commitdiff | tree
2015-07-10 James MolloyAdd support for fast-math flags to the FCmp instruction.
 commit | commitdiff | tree
2015-06-04 James MolloyDon't create a MIN/MAX node if the underlying compare...
commit | commitdiff | tree
2015-05-20 James MolloyReapply r237539 with a fix for the Chromium build.
 commit | commitdiff | tree
2015-05-17 James MolloyReapply r237520 with another fix for infinite looping
 commit | commitdiff | tree
2015-05-16 James MolloyRevert commits r237521 and r237520.
 commit | commitdiff | tree
2015-05-16 James MolloyUpdate to r237520 - swap order of CHECK-NEXT lines.
 commit | commitdiff | tree
2015-05-16 James MolloyReapply r237453 with a fix for the test timeouts.
 commit | commitdiff | tree
2015-05-15 James MolloyRevert "Canonicalize min/max expressions correctly."
 commit | commitdiff | tree
2015-05-15 James Molloy[SDAGBuilder] Make the AArch64 builder happier.
 commit | commitdiff | tree
2015-05-15 James MolloyMark SMIN/SMAX/UMIN/UMAX nodes as legal and add patterns...
commit | commitdiff | tree
2015-05-15 James MolloyCanonicalize min/max expressions correctly.
 commit | commitdiff | tree
2015-05-15 James MolloyAllow min/max detection to see through casts.
 commit | commitdiff | tree
2015-05-15 James Molloy[DependenceAnalysis] Fix for PR21585: collectUpperBound...
commit | commitdiff | tree
2015-05-15 James MolloyAdd SDNodes for umin, umax, smin and smax.
 commit | commitdiff | tree
2015-05-11 James MolloyRip min/max pattern matching out of InstCombine and...
commit | commitdiff | tree
2015-04-22 James Molloy[AArch64] Disable complex GEP optimization by default.
 commit | commitdiff | tree
2015-04-17 James MolloyFix TRUNCATE splitting helper logic.
 commit | commitdiff | tree
2015-04-16 James Molloy[AArch64] Fix invalid use of references to BuildMI.
 commit | commitdiff | tree
2015-03-31 James Molloy[SDAG] Move TRUNCATE splitting logic into a helper...
commit | commitdiff | tree
2015-03-27 James MolloyReapply r233175 and r233183: float2int.
 commit | commitdiff | tree
2015-03-25 James MolloyReapply r233062: "float2int": Add a new pass to demote...
commit | commitdiff | tree
2015-03-24 James Molloy"float2int": Add a new pass to demote from float to...
commit | commitdiff | tree
2015-03-23 James Molloy[ARM] Remove target-specific ITOFP/FPTOI nodes
 commit | commitdiff | tree
2015-03-06 James Molloy[ConstantRange] Teach multiply to be cleverer about...
commit | commitdiff | tree
2015-02-16 James MolloyRun LICM as part of the cleanup phase from the scalar...
commit | commitdiff | tree
2015-02-16 James Molloy[LoopReroll] Relax some assumptions a little.
 commit | commitdiff | tree
2015-02-16 James Molloy[LoopReroll] Don't crash on dead code
 commit | commitdiff | tree
2015-02-13 James Molloy[SimplifyCFG] Add test for r229099
 commit | commitdiff | tree
2015-02-13 James Molloy[SimplifyCFG] Be more aggressive
 commit | commitdiff | tree
2015-02-12 James Molloy[LoopRerolling] Be more forgiving with instruction...
commit | commitdiff | tree
2015-02-11 James MolloyMake buildbots better.
 commit | commitdiff | tree
2015-02-11 James Molloy[SimplifyCFG] Swap to using TargetTransformInfo for...
commit | commitdiff | tree
2015-02-11 James Molloy[LoopReroll] Introduce the concept of DAGRootSets.
 commit | commitdiff | tree
2015-01-29 James Molloy[LoopReroll] Alter the data structures used during...
commit | commitdiff | tree
2015-01-29 James Molloy[LoopReroll] Refactor most of reroll() into a helper...
commit | commitdiff | tree
2014-10-17 James Molloy[AArch64] Fix a silent codegen fault in BUILD_VECTOR...
commit | commitdiff | tree
2014-10-03 James MolloyRevert r215343.
 commit | commitdiff | tree
2014-09-27 James Molloy[AArch64] Redundant store instructions should be removed...
commit | commitdiff | tree
2014-09-14 James Molloy[A57FPLoadBalancing] Modify r217689 - actually we do...
commit | commitdiff | tree
2014-09-12 James Molloy[A57FPLoadBalancing] Remove support for vector types
 commit | commitdiff | tree
2014-09-12 James Molloy[A57FPLoadBalancing] Ignore <def>s when checking if...
commit | commitdiff | tree
2014-09-12 James Molloy[A57LoadBalancing] unique_ptr-ify.
 commit | commitdiff | tree
next
C/C++ LLVM-based compilers that forbids OOTA behaviors