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.

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

Unify device_memory_limit parsing and set default to 0.8 #439

Merged

Conversation

pentschev
Copy link
Member

This should address the request to set a default value for device_memory_limit, as well as allowing fractions and unifying how LocalCUDACluster and dask-cuda-worker parse its value. Fixes #334 .

cc @beckernick @rjzamora

@pentschev pentschev requested a review from a team as a code owner November 6, 2020 20:32
Copy link
Member

@quasiben quasiben left a comment

Choose a reason for hiding this comment

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

Thank you @pentschev . Should we also update documentation ?

@quasiben
Copy link
Member

quasiben commented Nov 6, 2020

I mean, do you want to call out the default setting in the docs? I'm happy to do this if you'd like

@pentschev
Copy link
Member Author

I mean, do you want to call out the default setting in the docs? I'm happy to do this if you'd like

Are you referring to some docstrings or https://dask-cuda.readthedocs.io/en/latest/ ?

@quasiben
Copy link
Member

quasiben commented Nov 6, 2020

@pentschev
Copy link
Member Author

Sure, we can update that next week. I just found out that we can't mix types with click.option, which causes problems with having float/int/str in --device-memory-limit, something that we do in LocalCUDACluster, so I'll have to find a solution for that too.

@pentschev
Copy link
Member Author

Sure, we can update that next week. I just found out that we can't mix types with click.option, which causes problems with having float/int/str in --device-memory-limit, something that we do in LocalCUDACluster, so I'll have to find a solution for that too.

Hopefully this is fixed in latest commits.

@codecov-io
Copy link

codecov-io commented Nov 6, 2020

Codecov Report

Merging #439 (6e51a2a) into branch-0.17 (3cefcc0) will increase coverage by 0.21%.
The diff coverage is 100.00%.

Impacted file tree graph

@@               Coverage Diff               @@
##           branch-0.17     #439      +/-   ##
===============================================
+ Coverage        90.57%   90.79%   +0.21%     
===============================================
  Files               14       14              
  Lines              796      804       +8     
===============================================
+ Hits               721      730       +9     
+ Misses              75       74       -1     
Impacted Files Coverage Δ
dask_cuda/cli/dask_cuda_worker.py 96.77% <ø> (ø)
dask_cuda/cuda_worker.py 71.95% <ø> (ø)
dask_cuda/local_cuda_cluster.py 81.33% <100.00%> (+0.32%) ⬆️
dask_cuda/utils.py 90.65% <100.00%> (+0.65%) ⬆️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 3cefcc0...6e51a2a. Read the comment docs.

@pentschev
Copy link
Member Author

@quasiben I updated the docs in 6e51a2a, please take a look when you have a chance.

@quasiben
Copy link
Member

quasiben commented Nov 9, 2020

Thanks @pentschev !

@quasiben quasiben merged commit e9aee6c into rapidsai:branch-0.17 Nov 9, 2020
@pentschev pentschev deleted the device-memory-limit-default-0.8 branch November 9, 2020 20:10
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.

[FEA] Allow spilling from device memory to host memory by default
3 participants