Skip to content

Betweenness Centrality Memory Allocation #5242

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
Draft
wants to merge 80 commits into
base: branch-25.10
Choose a base branch
from
Draft

Betweenness Centrality Memory Allocation #5242

wants to merge 80 commits into from

Conversation

HowardHuang1
Copy link
Contributor

No description provided.

@HowardHuang1
cpp implementation of brandes algo
018b9d5
@HowardHuang1
loop through frontier levels instead of all vertices --> small ~1.7x …
e59aa23 
…speedup
@HowardHuang1
new approach: distances extracted properly but still results in wrong…
d8a95fb 
… centrality values
@HowardHuang1
add debug statements for mismatched index issue
20da533
@HowardHuang1
Merge branch 'branch-25.08' of https://github.com/rapidsai/cugraph in…
bc0f2ef 
…to HH-betweenness-centrality
@HowardHuang1
Delete cpp/src/centrality/BC_pseudo.cuh
b8ae34b
@HowardHuang1
remove debug statements
566534b
@HowardHuang1
remove num_edges_to_process variable which was only used for the debu…
a3469e1 
…g statements that were removed
@HowardHuang1
per_v_transform_reduce_outgoing_e() API has restriction that it takes…
d374d2e 
… input array of size vertex_list and outputs array of same size vertex_list + use scatter to map array of size vertex_list back into global indices --> this fixed indexing issue
@HowardHuang1
clean up debug statements
6f3e0dc
@HowardHuang1
add test files for cugraph, nx-cugraph, and pylibcugraph
36681fe
@HowardHuang1
add road network test cases for libcugraph cpp tests
42f485b
@HowardHuang1
remove datasets
986292b
@HowardHuang1
move redundant computation of distance levels at every step out of th…
e68edd4 
…e loop
@HowardHuang1
add new york and california road network tests
ddfea92
@HowardHuang1
clean up debug statements
9ee19a0
@HowardHuang1
avoid scanning entire vertex range in every loop instead perform edge…
3c2bd46 
… and delta updates on frontier vertices only
@HowardHuang1
condense code with reusable buffers for vertices, deltas, and central…
de7e522 
…ities
@HowardHuang1
sort vertices by distance + use binary search to find lower bound and…
c7d7fdc 
… upper bound for each vertex frontier
@HowardHuang1
combine 3 thrust calls into 1
b1c457a
@HowardHuang1
time each chunk of code
d8b14bf
@HowardHuang1
move binary search for vertex frontier outside main diameter loop to …
1978c18 
…reduce kernel launches
@HowardHuang1
merge thrust calls to reduce kernel launches from 7 to 4 within each …
52e23ce 
…iteration of the delta update loop
@HowardHuang1
clean up timers and debug statements
00077e8
@HowardHuang1
simplify bound computation for binary search
664806a
@HowardHuang1
separate out combined thrust calls since it doesn't work for multi-gp…
e39f925 
…u case
@HowardHuang1
Merge remote-tracking branch 'upstream/branch-25.08' into HH-betweenn…
b707c48 
…ess-centrality
@HowardHuang1
add multi-gpu support
32cac28
@HowardHuang1
separate out distances, sigmas, and deltas in edge src/dst property
ccbd1a3
@HowardHuang1
Remove old files
ed697ca
@HowardHuang1
add timing for bfs
f42fe80
@HowardHuang1
revert commented out test cases and remove road network dataset
22680ca
@HowardHuang1
Merge remote-tracking branch 'upstream/branch-25.10' into HH-multisou…
adc3d35 
…rce-backwardpass
@HowardHuang1
add back in missed merge change
1ce46b0
@HowardHuang1
clean up timers and debug statements
2ffe91a
@HowardHuang1
combine delta and centrality for_each calls
66ed660
@HowardHuang1
eliminate count_if thrust call
4aede6d
@HowardHuang1
reduce array count from 4 to 2 to save memory
395eaba
@HowardHuang1
fix indent
b8bac05
@HowardHuang1
fix indents
c711e57
@HowardHuang1
ran pre-commit formatting
ca91c6b
@HowardHuang1
Revert CMakeLists.txt modifications
7843c07
@HowardHuang1
Revert betweenness_centrality_test.cpp modifications
6dcdbd4
@HowardHuang1
move multisource_bfs function into detail namespace
f0a2b3d
@HowardHuang1
move multisource_backward_pass from public to details namespace
af7f8e6
@HowardHuang1
change count_if to unique_count and atomicAdd to atomic_ref + fetch_add
efaabe2
@HowardHuang1
get rid of temporaries and redundancy
ea94b47
@HowardHuang1
modify buffer in place rather than copy to temporaries
56114ec
@HowardHuang1
modify buffer in place for BFS forward pass
7424217
@HowardHuang1
use public function extract_transform_if_v_frontier_outgoing_e instea…
abc0104 
…d of detail function
@HowardHuang1
use atomic_ref + fetch_add with the correct synchronization
443ca61
@HowardHuang1
default to sequential single source version for multi-gpu
59f9ea8
@HowardHuang1
fix minor syntax issues
efeab9b
@HowardHuang1
Merge 2 sorts into 1
730f6a8
@HowardHuang1
Merge remote-tracking branch 'upstream/branch-25.10' into HH-multisou…
257375a 
…rce-backwardpass
@HowardHuang1
add 4 stage memory allocation implementation
654bd63
@HowardHuang1
implement binary search memory allocation
d4bb679
@HowardHuang1
add env vars for fixed/dynamic mode and a user input field for fixed …
aa9c801 
…batch size
@HowardHuang1
start with 500
e7bd0c7
@copy-pr-bot copy-pr-bot
Copy link

copy-pr-bot bot commented Aug 28, 2025

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
1 participant
@HowardHuang1