Skip to content

Conversation

@cjac
Copy link
Contributor

@cjac cjac commented Jul 21, 2024

  • adding functions to check running environment OS
  • set rapids version to 23.12
  • set cuda version to 11.8
  • remove cudf variables - they are unused
  • set xgboost version to 2.0.3
  • created variable to store scala version for future move from 2.12 to 2.13
  • setting python_ver by operating system

cjac added 2 commits July 21, 2024 16:23
* adding functions to check running environment OS
* set rapids version to 23.12
* set cuda version to 11.8
* set cudf version - is this version used?
* set xgboost version to 2.0.3
* created variable to store scala version
* setting python_ver by operating system
* added functions to check for ubuntu variants
* removed definitions of unused variables having to do with CUDF
* capping max version for CUDA_VERSION to 11.8 which 12.4 implements probably
* capping max version of python to 3.10 until the stack works with python 3.11
@cjac
Copy link
Contributor Author

cjac commented Jul 22, 2024

Manually tested on

  • 2.0-ubuntu18 — ✅
  • 2.1-ubuntu20 — ✅
  • 2.2-ubuntu22 — ✅
  • 2.0-rocky8 — ✅
  • 2.2-rocky9 — ✅
  • 2.0-debian10 — ✅
  • 2.1-debian11 — ✅
  • 2.2-debian12 — ✅

@cjac cjac self-assigned this Jul 22, 2024
@cjac
Copy link
Contributor Author

cjac commented Jul 22, 2024

/gcbrun

@cjac
Copy link
Contributor Author

cjac commented Jul 22, 2024

/gcbrun

@cjac cjac force-pushed the rapids-20240721 branch from 889f5d8 to 1210097 Compare July 22, 2024 05:24
@cjac
Copy link
Contributor Author

cjac commented Jul 22, 2024

/gcbrun

@cjac
Copy link
Contributor Author

cjac commented Jul 22, 2024

this may depend on #1200 since I developed those changes as I was exercising these manual test runs.

@prince-cs
Copy link
Collaborator

Step #5 - "dataproc-2.1-debian11-tests": 2024-07-22T14:53:01.119357965Z NameError: name 'metadata' is not defined

https://github.com/GoogleCloudDataproc/initialization-actions/blob/master/rapids/test_rapids.py#L98

The presubmits seems to be failing because of this.

@prince-cs
Copy link
Collaborator

Also the timeout for presubmits needs to be increase to at least 4 hours.

readonly DEFAULT_CUDF_VERSION="22.10.0"
readonly DEFAULT_XGBOOST_VERSION="1.6.2"
readonly DEFAULT_XGBOOST_GPU_SUB_VERSION="0.3.0"
readonly DEFAULT_CUDA_VERSION="11.8"
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In GPU Driver we set the default cuda version to be 12.4. But, here it is 11.8. Any specific reason?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

11.8 is the maximum supported by mamba package, 'cudatools', which is ultimately the goal of this script.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Line no. 99: metadata is not required and is also not defined. Therefore, breaking the test flow.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It looks like it is required and must have been left out in a previous commit. fixed maybe. I'll gcbrun to see.

@cjac
Copy link
Contributor Author

cjac commented Jul 22, 2024

/gcbrun

2 similar comments
@prince-cs
Copy link
Collaborator

/gcbrun

@prince-cs
Copy link
Collaborator

/gcbrun

if self.getImageOs() == 'rocky':
self.skipTest("Not supported in Rocky Linux-based images")
# if self.getImageOs() == 'rocky':
# self.skipTest("Not supported in Rocky Linux-based images")
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

please just delete these.

def test_rapids_spark(self, configuration, machine_suffixes, accelerator):
if self.getImageOs() == 'rocky':
self.skipTest("Not supported in Rocky Linux-based images")
# if self.getImageOs() == 'rocky':
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ditto


if self.getImageOs() == 'rocky':
self.skipTest("Not supported in Rocky Linux-based images")
# if self.getImageOs() == 'rocky':
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ditto

@cjac cjac merged commit 40f33ce into GoogleCloudDataproc:master Jul 24, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants