1. 29 Nov, 2016 4 commits
  2. 28 Nov, 2016 3 commits
  3. 25 Nov, 2016 2 commits
  4. 24 Nov, 2016 3 commits
  5. 23 Nov, 2016 1 commit
  6. 22 Nov, 2016 5 commits
  7. 21 Nov, 2016 6 commits
  8. 19 Nov, 2016 1 commit
  9. 18 Nov, 2016 1 commit
  10. 17 Nov, 2016 7 commits
  11. 16 Nov, 2016 7 commits
    • Guillaume Pasero's avatar
      MRG: synchronize with branch 'develop', into bugfix-1278 · 86a986e0
      Guillaume Pasero authored
      Conflicts:
      	Modules/Visualization/MonteverdiGui/include/mvdImageViewRenderer.h
      86a986e0
    • Guillaume Pasero's avatar
      e9d65ef7
    • Guillaume Pasero's avatar
    • Guillaume Pasero's avatar
    • Julien Michel's avatar
    • Julien Michel's avatar
      PERF: Optimize BCOInterpolateImageFunction (kindly provided by Laurentiu Nicola) · b3a50337
      Julien Michel authored
      Simplify the code and avoid some allocations to improve performance.
      
      I haven't checked the definition of RealType, but merging the output and
      value variables might arguably affect precision. The unit tests still
      seem to pass, however.
      
      Tested by 2x upsampling a 10-band 5490x5490 float32 image on an Intel
      Celeron J1900.
      
      There is further room for improvement by eliding the remaining memory
      allocations. A container like the small_vector from Boost might be
      useful here, but we can't use that since it's not available in older
      Boost versions.
      
      before:
      661.44s user 9.31s system 304% cpu 3:40.52 total
      594.06s user 9.43s system 303% cpu 3:19.00 total
      591.15s user 9.43s system 301% cpu 3:19.28 total
      
      after:
      443.42s user 10.48s system 294% cpu 2:34.00 total
      443.11s user 9.87s system 309% cpu 2:26.27 total
      448.50s user 10.50s system 311% cpu 2:27.52 total
      b3a50337
    • Julien Michel's avatar
      PERF: Optimize BCOInterpolateImageFunction · 6a8c2a84
      Julien Michel authored
      Simplify the code and avoid some allocations to improve performance.
      
      I haven't checked the definition of RealType, but merging the output and
      value variables might arguably affect precision. The unit tests still
      seem to pass, however.
      
      Tested by 2x upsampling a 10-band 5490x5490 float32 image on an Intel
      Celeron J1900.
      
      There is further room for improvement by eliding the remaining memory
      allocations. A container like the small_vector from Boost might be
      useful here, but we can't use that since it's not available in older
      Boost versions.
      
      before:
      661.44s user 9.31s system 304% cpu 3:40.52 total
      594.06s user 9.43s system 303% cpu 3:19.00 total
      591.15s user 9.43s system 301% cpu 3:19.28 total
      
      after:
      443.42s user 10.48s system 294% cpu 2:34.00 total
      443.11s user 9.87s system 309% cpu 2:26.27 total
      448.50s user 10.50s system 311% cpu 2:27.52 total
      6a8c2a84