Code

Use CUDA math wheels#5966

Code

Merged

rapids-bot[bot] merged 12 commits intorapidsai:branch-24.10rapidsai/cuml:branch-24.10from KyleFromNVIDIA:dynamic-cuda-librariesKyleFromNVIDIA/cuml:dynamic-cuda-librariesCopy head branch name to clipboardAug 15, 2024

+71-4

Commits

Commits on Jul 16, 2024

• Use dynamic CUDA libraries

  KyleFromNVIDIA

  committed
  396ace6

  Copy full SHA for 396ace6

• Merge branch 'branch-24.08' into dynamic-cuda-libraries

  KyleFromNVIDIA

  committed
  6e5c87e

  Copy full SHA for 6e5c87e

Commits on Jul 17, 2024

• Only use CUDA 12+ wheels

  KyleFromNVIDIA

  committed
  b2bbe72

  Copy full SHA for b2bbe72

• Only find CUDAToolkit if cuml needs to be built

  KyleFromNVIDIA

  committed
  561e864

  Copy full SHA for 561e864

Commits on Jul 23, 2024

• Add USE_CUDA_MATH_WHEELS option

  KyleFromNVIDIA

  committed
  5475ebd

  Copy full SHA for 5475ebd

Commits on Aug 12, 2024

• Merge branch 'branch-24.10' into dynamic-cuda-libraries

  KyleFromNVIDIA

  committed
  c325647

  Copy full SHA for c325647

• Set auditwheel --exclude args based on CUDA version

  KyleFromNVIDIA

  committed
  82a9610

  Copy full SHA for 82a9610

• Review feedback

  KyleFromNVIDIA

  committed
  1c7cfcc

  Copy full SHA for 1c7cfcc

Commits on Aug 15, 2024

• Review feedback

  KyleFromNVIDIA

  committed
  656c9c9

  Copy full SHA for 656c9c9

• Specify USE_CUDA_MATH_WHEELS

  KyleFromNVIDIA

  committed
  e7b5158

  Copy full SHA for e7b5158

• USE_CUDA_MATH_WHEELS is off by default

  KyleFromNVIDIA

  committed
  ebe470f

  Copy full SHA for ebe470f

• Add nvjitlink back to rpath

  KyleFromNVIDIA

  committed
  d42059d

  Copy full SHA for d42059d