Commits

Branch selector

 master

User selector

d0k

Datepicker

All time

Commit History

Commits on Jul 7, 2014

• Make helper functions static.

  Show description for 0e278ad
  d0k

  committed
  0e278ad

  Copy full SHA for 0e278ad

• InstCombine: Simplify code, no functionality change.

  Show description for 8632e40
  d0k

  committed
  8632e40

  Copy full SHA for 8632e40

Commits on Jul 4, 2014

• GlobalDCE: Delete available_externally initializers if it allows removing the value the initializer is referring to.

  Show description for 5b55a47
  d0k

  committed
  5b55a47

  Copy full SHA for 5b55a47

• InstCombine: Strength reduce sadd.with.overflow into a regular nsw add if we can prove that it cannot overflow.

  Show description for 9c1df91
  d0k

  committed
  9c1df91

  Copy full SHA for 9c1df91

Commits on Jul 2, 2014

• X86: When combining shuffles just remove shuffles that are completely redundant.

  Show description for 11be760
  d0k

  committed
  11be760

  Copy full SHA for 11be760

Commits on Jul 1, 2014

• Remove unused typedef. GCC warns about this.

  Show description for 58d452b
  d0k

  committed
  58d452b

  Copy full SHA for 58d452b

• Rephrase loop so it doesn't leave unused bools around in Release mode.

  Show description for 4232f4e
  d0k

  committed
  4232f4e

  Copy full SHA for 4232f4e

Commits on Jun 24, 2014

• InstCombine: Disable umul.with.overflow recognition for vectors.

  Show description for 0e6156a
  d0k

  committed
  0e6156a

  Copy full SHA for 0e6156a

• InstCombine: Don't try to reorder shuffles where the mask is a ConstantExpr.

  Show description for 9c88403
  d0k

  committed
  9c88403

  Copy full SHA for 9c88403

Commits on Jun 22, 2014

• Add a description to the test from r211433 explaining why it's written that way.

  Show description for 84dd75d
  d0k

  committed
  84dd75d

  Copy full SHA for 84dd75d

Commits on Jun 21, 2014

• LoopUnrollRuntime: Check for overflow in the trip count calculation.

  Show description for 6b7ff6b
  d0k

  committed
  6b7ff6b

  Copy full SHA for 6b7ff6b

• Legalizer: Add support for splitting insert_subvectors.

  Show description for 636a9be
  d0k

  committed
  636a9be

  Copy full SHA for 636a9be

• SCEVExpander: Fold constant PHIs harder. The logic below only understands proper IVs.

  Show description for b7f1fb4
  d0k

  committed
  b7f1fb4

  Copy full SHA for b7f1fb4

Commits on Jun 6, 2014

• X86: Don't turn shifts into ands if there's another use that may not check for equality.

  Show description for 0da5960
  d0k

  committed
  0da5960

  Copy full SHA for 0da5960

Commits on May 31, 2014

• [Reassociate] Similar to "X + -X" -> "0", added code to handle "X + ~X" -> "-1".

  Show description for 37bc2f3
  d0k

  committed
  37bc2f3

  Copy full SHA for 37bc2f3

Commits on May 24, 2014

• MachineVerifier: Clean up some syntactic weirdness left behind by find&replace.

  Show description for c734d38
  d0k

  committed
  c734d38

  Copy full SHA for c734d38

• CodeGen: Make MachineBasicBlock::back skip to the beginning of the last bundle.

  Show description for 9d934ac
  d0k

  committed
  9d934ac

  Copy full SHA for 9d934ac

Commits on May 20, 2014

• Legalizer: Make bswap promotion safe for vectors.

  Show description for 51af588
  d0k

  committed
  51af588

  Copy full SHA for 51af588

Commits on May 19, 2014

• Flip on vectorization of bswap intrinsics.

  Show description for 29e068b
  d0k

  committed
  29e068b

  Copy full SHA for 29e068b

• SDAG: Legalize vector BSWAP into a shuffle if the shuffle is legal but the bswap not.

  Show description for bb81d9d
  d0k

  committed
  bb81d9d

  Copy full SHA for bb81d9d

Commits on May 18, 2014

• Options: Use erase_if to remove Args from the list.

  Show description for aad181f
  d0k

  committed
  aad181f

  Copy full SHA for aad181f

Commits on May 14, 2014

• X86: If we have an instruction that sets a flag and a zero test on the input of that instruction try to eliminate the test.

  Show description for 202be06
  d0k

  committed
  202be06

  Copy full SHA for 202be06

Commits on May 13, 2014

• GVN: Fix non-determinism in map iteration.

  Show description for 5b6887d
  d0k

  committed
  5b6887d

  Copy full SHA for 5b6887d

• GVN: rangify a couple of loops.

  Show description for 7fe69dd
  d0k

  committed
  7fe69dd

  Copy full SHA for 7fe69dd

Commits on May 12, 2014

• X86: Make sure that we have SSE4.1 before we generate insertps nodes.

  Show description for b31a977
  d0k

  committed
  b31a977

  Copy full SHA for b31a977

Commits on May 11, 2014

• SLPVectorizer: Instead of just performing CSE on dead blocks ignore them completely.

  Show description for 4db3f41
  d0k

  committed
  4db3f41

  Copy full SHA for 4db3f41

Commits on May 10, 2014

• SCEV: Use range-based for loop and fold variable into assert.

  Show description for bfd5dad
  d0k

  committed
  bfd5dad

  Copy full SHA for bfd5dad

Commits on May 9, 2014

• SLPVectorizer: When sorting by domination for CSE don't assert on unreachable code.

  Show description for 3ae331d
  d0k

  committed
  3ae331d

  Copy full SHA for 3ae331d

• [asan] Stop leaking X86Operands.

  Show description for d8d1a7c
  d0k

  committed
  d8d1a7c

  Copy full SHA for d8d1a7c

Commits on May 6, 2014

• TTI: Estimate @llvm.fmuladd cost as fmul + fadd when FMA's aren't legal on the target.

  Show description for 2c06cd8
  d0k

  committed
  2c06cd8

  Copy full SHA for 2c06cd8

Commits on May 4, 2014

• LoopUnroll: If we're doing partial unrolling, use the PartialThreshold to limit unrolling.

  Show description for 99b03e3
  d0k

  committed
  99b03e3

  Copy full SHA for 99b03e3

Commits on May 3, 2014

• SLPVectorizer: Lazily allocate the map for block numbering.

  Show description for 81f28f6
  d0k

  committed
  81f28f6

  Copy full SHA for 81f28f6

Commits on May 2, 2014

• Satisfy GCC's urgent need for parentheses around ‘&&’ within ‘||’.

  Show description for 71aafce
  d0k

  committed
  71aafce

  Copy full SHA for 71aafce

• Add a description for AMD's bdver4 (aka Excavator).

  Show description for 3cddd16
  d0k

  committed
  3cddd16

  Copy full SHA for 3cddd16

• Allow SelectionDAG::FoldConstantArithmetic to work when it's called with a vector VT but scalar values.

  Show description for bcf1501
  d0k

  committed
  bcf1501

  Copy full SHA for bcf1501

Pagination

Previous

Next