Code

Optimize compaction operations#10030

Code

Merged

rapids-bot[bot] merged 55 commits intorapidsai:branch-22.04rapidsai/cudf:branch-22.04from PointKernel:optimize-compactionPointKernel/cudf:optimize-compactionCopy head branch name to clipboardFeb 2, 2022

+1,170-488

Commits

Commits on Jan 12, 2022

• Rename existing compaction APIs

  PointKernel

  committed
  2f33e04

  Copy full SHA for 2f33e04

• Merge remote-tracking branch 'upstream/branch-22.02' into optimize-compaction

  PointKernel

  committed
  3212706

  Copy full SHA for 3212706

• Update cython code to accommodate renaming

  PointKernel

  committed
  7ce9549

  Copy full SHA for 7ce9549

• Update copyrights

  PointKernel

  committed
  5c5a415

  Copy full SHA for 5c5a415

• Refactor unordered_distinct_count with hash-based algorithms

  PointKernel

  committed
  0a62ade

  Copy full SHA for 0a62ade

• Merge remote-tracking branch 'upstream/branch-22.02' into optimize-compaction

  PointKernel

  committed
  e6acd8f

  Copy full SHA for e6acd8f

Commits on Jan 13, 2022

• Refactor unordered_drop_duplicates with hash-based algorithms

  PointKernel

  committed
  fba851c

  Copy full SHA for fba851c

• Update cython code

  PointKernel

  committed
  05ee85f

  Copy full SHA for 05ee85f

• Optimize distinct count: insert valid rows only if nulls are equal

  PointKernel

  committed
  8ab22a4

  Copy full SHA for 8ab22a4

• Fill column via mutable view + update comments

  PointKernel

  committed
  bba7b57

  Copy full SHA for bba7b57

• Minor corrections

  PointKernel

  committed
  6746f28

  Copy full SHA for 6746f28

• Update benchmarks and unit tests

  PointKernel

  committed
  70292bc

  Copy full SHA for 70292bc

• Add reminder for further optimization in distinct count

  PointKernel

  committed
  46d83b9

  Copy full SHA for 46d83b9

• Fix transform test failure

  PointKernel

  committed
  f07d3d0

  Copy full SHA for f07d3d0

• Fix dictionary test failures

  PointKernel

  committed
  8fcfbae

  Copy full SHA for 8fcfbae

• Merge remote-tracking branch 'upstream/branch-22.04' into optimize-compaction

  PointKernel

  committed
  a8fe478

  Copy full SHA for a8fe478

Commits on Jan 14, 2022

• Add sort-based implementations back to the repo

  PointKernel

  committed
  f2ac25d

  Copy full SHA for f2ac25d

• Update copyright

  PointKernel

  committed
  0ed5712

  Copy full SHA for 0ed5712

• Add consecutive distinct_count

  PointKernel

  committed
  e372144

  Copy full SHA for e372144

Commits on Jan 15, 2022

• Remove nan control in distinct_count

  PointKernel

  committed
  fc57b29

  Copy full SHA for fc57b29

• Update unit tests

  PointKernel

  committed
  d810e0b

  Copy full SHA for d810e0b

Commits on Jan 17, 2022

• Add nan handling to distinct_count + update unit tests

  PointKernel

  committed
  40cc410

  Copy full SHA for 40cc410

• Rename drop_duplicates as sort_and_drop_duplicates

  PointKernel

  committed
  dd91e64

  Copy full SHA for dd91e64

• Add consecutive drop_duplicates

  PointKernel

  committed
  d80911c

  Copy full SHA for d80911c

• Optimize unordered_distinct_count: insert non-null rows only to improve efficiency

  PointKernel

  committed
  7ced995

  Copy full SHA for 7ced995

• Update cuco git tag

  PointKernel

  committed
  2ea5d8e

  Copy full SHA for 2ea5d8e

Commits on Jan 18, 2022

• Slience unused argument warning via function prototyping

  PointKernel

  committed
  4bb7b16

  Copy full SHA for 4bb7b16

• Refactor compaction benchmark with nvbench

  PointKernel

  committed
  012ca8b

  Copy full SHA for 012ca8b

• Update copyright

  PointKernel

  committed
  d489e2e

  Copy full SHA for d489e2e

Commits on Jan 19, 2022

• Get rid of nvbench primitive types

  PointKernel

  committed
  3e47ffd

  Copy full SHA for 3e47ffd

• Update docs & comments

  PointKernel

  committed
  a0a10e5

  Copy full SHA for a0a10e5

• Address review comments

  PointKernel

  committed
  5fb92c7

  Copy full SHA for 5fb92c7

• Address more review comments

  PointKernel

  committed
  3af4fd0

  Copy full SHA for 3af4fd0

• Split tests

  PointKernel

  committed
  a587511

  Copy full SHA for a587511

• Use null masks in tests

  PointKernel

  committed
  b062eb5

  Copy full SHA for b062eb5

• Split benchmarks

  PointKernel

  committed
  a5f881f

  Copy full SHA for a5f881f

Commits on Jan 21, 2022

• Fix a bug + update tests

  PointKernel

  committed
  df36e77

  Copy full SHA for df36e77

• Update docs

  PointKernel

  committed
  20ed6ea

  Copy full SHA for 20ed6ea

• Merge remote-tracking branch 'upstream/branch-22.04' into optimize-compaction

  PointKernel

  committed
  e401690

  Copy full SHA for e401690

• Add should_check_nan predicate to avoid unnecessary type-dispatching

  PointKernel

  committed
  ecc1d7e

  Copy full SHA for ecc1d7e

• Rename benchmark according to benchmarking guide

  PointKernel

  committed
  a151443

  Copy full SHA for a151443

Commits on Jan 24, 2022

• Remove std::unique-like drop_duplicates

  PointKernel

  committed
  024d7e0

  Copy full SHA for 024d7e0

• Style fixing

  PointKernel

  committed
  b6c1634

  Copy full SHA for b6c1634

• Fix test failures: sort the output

  PointKernel

  committed
  3ad0f76

  Copy full SHA for 3ad0f76

• Minor cleanups

  PointKernel

  committed
  58f6cb6

  Copy full SHA for 58f6cb6

• Minor cleanup

  PointKernel

  committed
  e381815

  Copy full SHA for e381815

• Address review comments

  PointKernel

  committed
  fa796aa

  Copy full SHA for fa796aa

Commits on Jan 25, 2022

• Merge remote-tracking branch 'upstream/branch-22.04' into optimize-compaction

  PointKernel

  committed
  3915134

  Copy full SHA for 3915134

Commits on Jan 27, 2022

• Address review comments

  PointKernel

  committed
  118468e

  Copy full SHA for 118468e

• Simply if logic

  PointKernel

  committed
  d1535d5

  Copy full SHA for d1535d5

• Minor updates

  PointKernel

  committed
  0b0d015

  Copy full SHA for 0b0d015

• Add early exit

  PointKernel

  committed
  906f469

  Copy full SHA for 906f469

Commits on Jan 28, 2022

• Fix cuco pair issues with the latest cuco tag

  PointKernel

  committed
  c8a3e87

  Copy full SHA for c8a3e87

Commits on Feb 2, 2022

• Address review comments

  PointKernel

  committed
  070d5ce

  Copy full SHA for 070d5ce

• Address review + update comments

  PointKernel

  committed
  a60c128

  Copy full SHA for a60c128