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

Performance-optimize storing edge partition source/destination properties in (key, value) pairs #2328

Merged
merged 18 commits into from
Jun 7, 2022

Conversation

seunghwak
Copy link
Contributor

@seunghwak seunghwak commented Jun 2, 2022

Using (key, value) pairs instead of linear array for the entire source/destination range cuts memory footprint significantly when # GPUs >> (V/E)^2.

This PR performance-optimizes finding the value for the given key and also enables (key, value) pairs (this feature was previously disabled by default).

The original implementation binary searched over the entire set of local source/destination keys. This PR compared a hash table based approach (using cuCollection) and the auxiliary array based approach (described in the original DCSC paper) and selected the auxiliary array based approach.

Breaking due to few graph primitives API changes.

@seunghwak seunghwak added 2 - In Progress improvement Improvement / enhancement to an existing function breaking Breaking change labels Jun 2, 2022
@seunghwak seunghwak added this to the 22.08 milestone Jun 2, 2022
@seunghwak seunghwak self-assigned this Jun 2, 2022
@seunghwak seunghwak changed the title Performance-optimize storing edge partition source/destination properties in (key, value) pairs [skip-ci] Performance-optimize storing edge partition source/destination properties in (key, value) pairs Jun 3, 2022
@codecov-commenter
Copy link

codecov-commenter commented Jun 3, 2022

Codecov Report

❗ No coverage uploaded for pull request base (branch-22.08@27fceff). Click here to learn what that means.
The diff coverage is n/a.

@@               Coverage Diff               @@
##             branch-22.08    #2328   +/-   ##
===============================================
  Coverage                ?   60.74%           
===============================================
  Files                   ?      105           
  Lines                   ?     5075           
  Branches                ?        0           
===============================================
  Hits                    ?     3083           
  Misses                  ?     1992           
  Partials                ?        0           

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 27fceff...c265af2. Read the comment docs.

@seunghwak seunghwak changed the title [skip-ci] Performance-optimize storing edge partition source/destination properties in (key, value) pairs Performance-optimize storing edge partition source/destination properties in (key, value) pairs Jun 3, 2022
@seunghwak seunghwak marked this pull request as ready for review June 3, 2022 18:37
@seunghwak seunghwak requested a review from a team as a code owner June 3, 2022 18:37
@seunghwak seunghwak requested a review from kaatish June 3, 2022 18:41
static_cast<double>(num_local_unique_edge_minors) / static_cast<double>(minor_size),
raft::comms::op_t::MAX,
handle.get_stream());
std::cout << "max_major_properties_fill_ratio=" << max_major_properties_fill_ratio
Copy link
Collaborator

Choose a reason for hiding this comment

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

Residual debug statement?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Oops, I though I deleted this but it seems like I didn't :-) I will delete this.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Deleted.

@ChuckHastings
Copy link
Collaborator

@gpucibot merge

@rapids-bot rapids-bot bot merged commit 7100fd5 into rapidsai:branch-22.08 Jun 7, 2022
@seunghwak seunghwak deleted the enh_scaling branch August 11, 2022 23:25
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
breaking Breaking change improvement Improvement / enhancement to an existing function
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants