Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Evaluate dask-cuml for 0.8 #700

Closed
JohnZed opened this issue Jun 17, 2019 · 2 comments
Closed

Evaluate dask-cuml for 0.8 #700

JohnZed opened this issue Jun 17, 2019 · 2 comments
Assignees
@cjnolet
Labels
2 - In Progress Currenty a work in progress

Comments

@JohnZed
Copy link
Contributor

JohnZed commented Jun 17, 2019

No description provided.

@cjnolet cjnolet closed this as completed Jun 18, 2019
@cjnolet
Copy link
Member

cjnolet commented Jun 20, 2019

Opening this back up because pending CI issues are dragging on and I don't want to lose sight of this for the release.

So far I have verified locally w/ the libcumlmg that test_linear_regression.py and test_nearest_neighbors.py succeed when run.

Unfortunately, the following two issues arose in CI:

  • The random projection algorithm is doing a binomial sampling w/o replacement on host, which is pulling in a log gamma function that was built with a newer version of glibc than the one that comes w/ centos (centos has version 19, this was built w/ 23). This likely results from the use of a different libstdc++ lib when linking than what is used at runtime. We are still verifying but believe this is the main issue.
  • The most recent build of dask-cuda has an error which is causing the knn test to fail. This is upstream from us and, according to Matthew Rocklin, is upstream from dask-cuda as well.

Once we have confirmed it, I will make an issue #1 so that it can be fixed properly before the 0.8 release.

@cjnolet cjnolet reopened this Jun 20, 2019
@cjnolet cjnolet mentioned this issue Jun 20, 2019
Closed
@cjnolet cjnolet added the 2 - In Progress Currenty a work in progress label Jun 21, 2019
@cjnolet
Copy link
Member

cjnolet commented Jun 21, 2019

Aside from upstream problems in the dask-cuda project, this is good to go

@cjnolet cjnolet closed this as completed Jun 24, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@cjnolet cjnolet

Labels
2 - In Progress Currenty a work in progress
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@JohnZed @cjnolet