Code

Optimization pass#58

Code

Merged

ljubobratovicrelja merged 48 commits intomasterlibmir/dcv:masterfrom optimization-patch-2libmir/dcv:optimization-patch-2Copy head branch name to clipboardNov 12, 2016

+1,314-919

Commits

Commits on Sep 29, 2016

• Parallelized and optimized corner response calculation.

  ljubobratovicrelja

  committed
  d870aa6

Commits on Oct 1, 2016

• Refactoring of extractCorners.

  ljubobratovicrelja

  committed
  daf4e92

• Refactoring rgb(bgr)2gray.

  ljubobratovicrelja

  committed
  958f5a9

• Refactoring rgb2hsv.

  ljubobratovicrelja

  committed
  b6ac069

• Refactoring hsv2rgb.

  ljubobratovicrelja

  committed
  7da0159

• uninitializedArray --> uninitializedSlice.

  ljubobratovicrelja

  committed
  5658eb9

• Refactoring yuv2rgb, overall module cleanup.

  ljubobratovicrelja

  committed
  0fc5b01

Commits on Oct 2, 2016

• Added static slice packing helpers.

  ljubobratovicrelja

  committed
  5970cb2

• Extracted ndIter kernel lambdas to functions.

  ljubobratovicrelja

  committed
  720fd50

• div by 100.0 replaced with mul by 0.01.

  ljubobratovicrelja

  committed
  21f6118

• Extracted kernels to functions, added fastmath.

  ljubobratovicrelja

  committed
  c3cb076

• Replaced std.math.sqrt with llvm_sqrt.

  ljubobratovicrelja

  committed
  5611394

• Added mcpu=native flag.

  ljubobratovicrelja

  committed
  deaeed9

• Randomized hsv values.

  ljubobratovicrelja

  committed
  e5228ed

• Loop refactoring.

  ljubobratovicrelja

  committed
  aead0ad

• Fixed non-contiguous loop iteration fail.

  ljubobratovicrelja

  committed
  09b5f1f

Commits on Oct 6, 2016

• Added nothrow and nogc.

  ljubobratovicrelja

  committed
  444f13a

• Refactoring of convolution functions.

  ljubobratovicrelja

  committed
  e4075f1

Commits on Oct 9, 2016

• Merge branch 'master' into optimization-patch-2

  ljubobratovicrelja

  committed
  a486d08

Commits on Oct 10, 2016

• Merge branch 'master' into optimization-patch-2

  ljubobratovicrelja

  committed
  a58eb04

Commits on Oct 11, 2016

• Added tensor border extraction.

  ljubobratovicrelja

  committed
  7ab3a0b

• Added in contract for non-matrix tensors.

  ljubobratovicrelja

  committed
  f352a77

• Reorganized templates in boundary conditions to support lazy tensors.

  ljubobratovicrelja

  committed
  4a9722a

• Temporary silence of isBoundaryCondition.

  ljubobratovicrelja

  committed
  6a1023e

• Refactoring of bilateralFilter function.

  ljubobratovicrelja

  committed
  66367dc

• Tweaks in bilateralDisparityFilter to match bilateralFilter changes.

  ljubobratovicrelja

  committed
  56d1ede

• Tweaks in bilateralFilter tests to match bilateralFilter changes.

  ljubobratovicrelja

  committed
  a87f2bd

• Fixed multi-channel bilateralFilter.

  ljubobratovicrelja

  committed
  549cbab

• Reorganized convolution templates to support lazy tensors.

  ljubobratovicrelja

  committed
  539aced

• Adopted convolution changes in benchmark functions.

  ljubobratovicrelja

  committed
  13fc0b6

Commits on Oct 18, 2016

• Refactoring of calcGradient, calcPartialDerivatives.

  ljubobratovicrelja

  committed
  ecc4b0e

• Refactoring of nonMaximaSuppression.

  ljubobratovicrelja

  committed
  4569ffb

• Refactored scaled and ranged to accept lazy tensors.

  ljubobratovicrelja

  committed
  7f2f5b7

Commits on Oct 19, 2016

• Parallelized calcPartialDerivatives, calcGradients and canny.

  ljubobratovicrelja

  committed
  2cd6343

• Fixed docs typo.

  ljubobratovicrelja

  committed
  49a6963

Commits on Nov 3, 2016

• Merge branch 'master' into optimization-patch-2

  ljubobratovicrelja

  committed
  ba7b6ac

• Merge branch 'master' into optimization-patch-2

  ljubobratovicrelja

  committed
  1d2b8e4

Commits on Nov 4, 2016

• Refactoring of filterNonMaximum.

  ljubobratovicrelja

  committed
  b21c660

• Docs fixup.

  ljubobratovicrelja

  committed
  664a58f

• filterNonMaximum optimization.

  ljubobratovicrelja

  committed
  fcd1ede

Commits on Nov 5, 2016

• Added type check for filterNonMaximum input.

  ljubobratovicrelja

  committed
  b83ae5a

Commits on Nov 10, 2016

• Refactoring of warp/remap.

  ljubobratovicrelja

  committed
  c42e8fd

• Added fastmath on linear interpolation.

  ljubobratovicrelja

  committed
  0a83b39

• Updated function template arguments.

  ljubobratovicrelja

  committed
  24e8fbb

• Refactoring of threshold.

  ljubobratovicrelja

  committed
  0c744a7

Commits on Nov 11, 2016

• Fixed fastmath floor.

  ljubobratovicrelja

  committed
  7774f26

• Merge with master.

  ljubobratovicrelja

  committed
  824a4a8

Commits on Nov 12, 2016

• Docs and formatting, asserts and notes on contiguous slice request.

  ljubobratovicrelja

  committed
  d88c786