Commits

Branch selector

 master

User selector

milseman

Datepicker

All time

Commit History

Commits on Dec 5, 2013

• Use present fast-math flags when applicable in CreateBinOp

  Show description for faf4d59
  milseman

  committed
  faf4d59

Commits on Nov 19, 2013

• Add support for software expansion of 64-bit integer division instructions.

  Show description for 1b3ab91
  milseman

  committed
  1b3ab91

Commits on Mar 8, 2013

• Early exit from getAllocationData() and isFreeCall() for intrinsics.

  Show description for cacf971
  milseman

  committed
  cacf971

• Remove trailing whitespace

  Show description for 9333ffb
  milseman

  committed
  9333ffb

Commits on Mar 1, 2013

• Cache the result of Function::getIntrinsicID() in a DenseMap attached to the LLVMContext.

  Show description for 4c8e74f
  milseman

  committed
  4c8e74f

Commits on Feb 27, 2013

• Reverted: r176136 - Have a way for a target to opt-out of target-independent fast isel

  Show description for 7dbd34b
  milseman

  committed
  7dbd34b

Commits on Feb 26, 2013

• Have a way for a target to opt-out of target-independent fast isel

  Show description for bdc98d3
  milseman

  committed
  bdc98d3

• Constant fold vector bitcasts of halves similarly to how floats and doubles are folded. Test case included.

  Show description for 6160250
  milseman

  committed
  6160250

• Use a DenseMap instead of a std::map for AnalysisID -> Pass* maps. This reduces the pass-manager overhead from FPPassManager::runOnFunction() by about 10%.

  Show description for ce522ee
  milseman

  committed
  ce522ee

Commits on Feb 11, 2013

• Optimization: bitcast (<1 x ...> insertelement ..., X, ...) to ... ==> bitcast X to ...

  Show description for 9c213cc
  milseman

  committed
  9c213cc

• Remove trailing whitespace

  Show description for 651fb49
  milseman

  committed
  651fb49

Commits on Feb 7, 2013

• Have InstCombine call SipmlifyCall when handling calls. Test case included.

  Show description for 3854a5d
  milseman

  committed
  3854a5d

• Identify and simplify idempotent intrinsics. Test case included.

  Show description for f89de81
  milseman

  committed
  f89de81

• Preserve fast-math flags after reassociation and commutation. Update test cases

  Show description for 0fae64f
  milseman

  committed
  0fae64f

• whitespace

  Show description for 9f41a5c
  milseman

  committed
  9f41a5c

Commits on Jan 21, 2013

• Docs for SparseMultiSet

  Show description for e4b1efe
  milseman

  committed
  e4b1efe

• Introduce a new data structure, the SparseMultiSet, and changes to the MI scheduler to use it.

  Show description for afe77f3
  milseman

  committed
  afe77f3

Commits on Jan 11, 2013

• Support for half intrinsics. Pushes MMX into slower encoding path.

  Show description for 4d0b4a4
  milseman

  committed
  4d0b4a4

Commits on Dec 19, 2012

• Refactor isIntrinsic() to be quicker, and change classof() (and thus, isa<IntrinsicInst>()) to use it. This decreases the number of occurrences of the slow-path string matching performed by getIntr…

  Show description for f846f16
  milseman

  committed
  f846f16

Commits on Dec 17, 2012

• Remove trailing whitespace

  Show description for f3ceaff
  milseman

  committed
  f3ceaff

• Removed trailing whitespace

  Show description for 0f47e7e
  milseman

  committed
  0f47e7e

Commits on Dec 14, 2012

• Add back FoldOpIntoPhi optimizations with fix. Included test cases to help catch these errors and to test the presence of the optimization itself

  Show description for 07acee7
  milseman

  committed
  07acee7

Commits on Dec 13, 2012

• Pattern matching code for intrinsics.

  Show description for 8ad435f
  milseman

  committed
  8ad435f

• m_CombineOr and m_CombineAnd pattern combinators

  Show description for 8961924
  milseman

  committed
  8961924

Commits on Dec 12, 2012

• Restore the PHI optimization I accidently removed

  Show description for 94993e3
  milseman

  committed
  94993e3

• Remove trailing whitespace

  Show description for 4d96e6f
  milseman

  committed
  4d96e6f

• Have SimplifyBinOp call the new FAdd/FSub/FMul helpers, with fast-math flags off

  Show description for d0a0d22
  milseman

  committed
  d0a0d22

• Remove redunant optimizations from InstCombine, instead call the appropriate functions from SimplifyInstruction

  Show description for c244f38
  milseman

  committed
  c244f38

• Added a slew of SimplifyInstruction floating-point optimizations, many of which take advantage of fast-math flags. Test cases included.

  Show description for 09ee250
  milseman

  committed
  09ee250

• Pattern matchers for floating point values

  Show description for 1d4b150
  milseman

  committed
  1d4b150

• Remove FIXMEs surrounding Constant[Data]Vectors, instead

  Show description for 08b25ce
  milseman

  committed
  08b25ce

Commits on Dec 9, 2012

• Reorganize FastMathFlags to be a wrapper around unsigned, and streamline some interfaces.

  Show description for 1638b83
  milseman

  committed
  1638b83

• Have the bitcode reader/writer just use FPMathOperator's fast math enum directly

  Show description for 855d025
  milseman

  committed
  855d025

Commits on Dec 7, 2012

• Remove trailing whitespace

  Show description for 829c8bd
  milseman

  committed
  829c8bd

Commits on Dec 6, 2012

• Have CannotBeNegativeZero() be aware of the nsz fast-math flag

  Show description for 85893f4
  milseman

  committed
  85893f4

Pagination

PreviousNext