Merge latest rapidsai:branch-0.19 into aschaffer:branch-0.19 #44
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…hpp file. (#1476) Adds cmake function and .hpp template for generating a `version_config.hpp` file, similar to RMM's file of the same name. This allows C++ clients to include the file from the libcugraph install to query version information for reporting, checking compatibility, etc. Tested by building and installing libcugraph and checking that `version_config.hpp` was present in the conda environment and contained the correct information. closes #1472 FYI @anaruse Authors: - Rick Ratzel (@rlratzel) Approvers: - Brad Rees (@BradReesWork) - Alex Fender (@afender) URL: #1476
This provides fixes for strongly connected components on graphs with self-loops: #1471. closes #1471 Authors: - Andrei Schaffer (@aschaffer) Approvers: - Brad Rees (@BradReesWork) - Rick Ratzel (@rlratzel) URL: #1475
This updates the default conda build directory to a custom path for gpuCI. Small changes to artifact paths as well to make this compatible with Project Flash. Authors: - Dillon Cullinan (@dillon-cullinan) Approvers: - AJ Schmidt (@ajschmidt8) URL: #1452
Addressed comments from: #1450 In addition: 1. Made raft handle const. 2. Split algorithm into multiple kernel calls and updated timers. 3. Removed global symbols to retrieve results in struct. Authors: - Hugo Linsenmaier (@hlinsen) Approvers: - Brad Rees (@BradReesWork) - Andrei Schaffer (@aschaffer) - Alex Fender (@afender) URL: #1480
There are various things in this PR. Multi-Seed (MS) BFS: - API tentative - Saving research on memory feasability helper function (not in production) - Saving research on running the current BFS concurrently with streams and threads for analysis perf comparison (not in production) EgoNet: - Multithreading in EgoNet which deserializes execution and comes with mild performance improvements on large sizes - Some cleanup Authors: - Alex Fender (@afender) Approvers: - Chuck Hastings (@ChuckHastings) - @Iroy30 - Brad Rees (@BradReesWork) URL: #1469
Ensure all conda packages created in this repo that depend on other packages are all version pinned to the same build number. This way it prevents a conda solve from picking mismatched versions of `cugraph` and `libcugraph` that can break this repo during builds and testing. Authors: - Mike Wendt (@mike-wendt) Approvers: - Brad Rees (@BradReesWork) - Ray Douglass (@raydouglass) URL: #1485
…1423) Implement the `update_by_delta_modularity` method using the new graph primitives and pattern accelerators. This eliminates all of the custom MNMG implementation originally created for MNMG Louvain a few releases ago and replaces it with the new pattern accelerator and graph primitives that have been added in the last couple of releases. This depends on the following PRs and should not be merged until after them: * #1394 * #1399 closes #1220 Authors: - Chuck Hastings (@ChuckHastings) Approvers: - Andrei Schaffer (@aschaffer) - Seunghwa Kang (@seunghwak) - Rick Ratzel (@rlratzel) - Alex Fender (@afender) URL: #1423
Authors: - @Iroy30 Approvers: - Chuck Hastings (@ChuckHastings) - Alex Fender (@afender) - Seunghwa Kang (@seunghwak) URL: #1484
This PR tracks work on issue: #1380. Authors: - Andrei Schaffer (@aschaffer) Approvers: - Brad Rees (@BradReesWork) - Chuck Hastings (@ChuckHastings) - Seunghwa Kang (@seunghwak) URL: #1439
Add a function that leverages the RMAT generator and creates k graphs with a power-law or uniform distribution of graph sizes. Closes #1458 Consider adding bindings for this as part of #1473 Authors: - Alex Fender (https://github.com/afender) Approvers: - Seunghwa Kang (https://github.com/seunghwak) - Brad Rees (https://github.com/BradReesWork) URL: #1483
Improve the C++ documentation for the following algorithms: Pagerank, SSSP, BFS, and Louvain. We need, generally, to improve the C++ documentation for cuGraph. This is intended to improve the documentation for the algorithms that have been adapted to use the new graph primitives and will hopefully provide a construct that we can use for updating other C++ documentation. As we migrate new algorithms to use the new graph primitives over the next several releases we will create documentation for those algorithms as well. Closes #1490 Authors: - Chuck Hastings (https://github.com/ChuckHastings) Approvers: - Rick Ratzel (https://github.com/rlratzel) - Brad Rees (https://github.com/BradReesWork) URL: #1489
Reverts #1485 Authors: - Ray Douglass (https://github.com/raydouglass) Approvers: - Mike Wendt (https://github.com/mike-wendt) URL: #1493
#1498) * Made various fixes to `test_mg_betweenness_centrality.py` to address recent updates that were breaking these tests. * @afender updated egonet.cu to address an occasional test failure related to mem usage in certain environments. * @rlratzel updated python and notebook tests that contain code not supported on Pascal to be skipped when running on that arch. Also added a script to detect Pascal and used it to skip **all** C++ tests. _Note: a better way would be to edit each C++ test to use GTEST_SKIP to conditionally skip on Pascal, but since there are now so many to skip, and we get coverage from Python, this was the faster solution without sacrificing much/any coverage for a single platform._ Authors: - Joseph Nke (https://github.com/jnke2016) - Alex Fender (https://github.com/afender) - Rick Ratzel (https://github.com/rlratzel) Approvers: - Alex Fender (https://github.com/afender) - Rick Ratzel (https://github.com/rlratzel) - AJ Schmidt (https://github.com/ajschmidt8) URL: #1498
Cugraph fix for: #1499 Authors: - Hugo Linsenmaier (https://github.com/hlinsen) Approvers: - Brad Rees (https://github.com/BradReesWork) - Alex Fender (https://github.com/afender) URL: #1504
Authors: - https://github.com/Iroy30 - Rick Ratzel (https://github.com/rlratzel) Approvers: - Seunghwa Kang (https://github.com/seunghwak) - Alex Fender (https://github.com/afender) URL: #1505
By explicitly telling nvcc's fatbin pass to always compress device code we can ensure that our binaries are the smallest possible size. See rapidsai/cudf#7583 for additional context. Authors: - Robert Maynard (https://github.com/robertmaynard) Approvers: - Rick Ratzel (https://github.com/rlratzel) - Brad Rees (https://github.com/BradReesWork) URL: #1503
There was a PR merged into RAFT today, updating to the most recent commit hash : f0cd81f. Authors: - Alex Fender (https://github.com/afender) Approvers: - Seunghwa Kang (https://github.com/seunghwak) - Brad Rees (https://github.com/BradReesWork) URL: #1509
Partially addresses Issue #1442 Update graph partitioning scheme to better control memory footprint vs concurrency trade-offs for large-scale graph processing in large clusters. This new partitioning scheme also simplifies communication patterns among GPUs which can potentially improve scalability. Authors: - Seunghwa Kang (https://github.com/seunghwak) Approvers: - Chuck Hastings (https://github.com/ChuckHastings) - Alex Fender (https://github.com/afender) - Andrei Schaffer (https://github.com/aschaffer) URL: #1443
Updated NetworkX version to latest version, which addresses an incompatibility with the latest `decorator` dependency. Tested by running the BC tests which were previously failing with Nx 2.5 Authors: - Rick Ratzel (https://github.com/rlratzel) Approvers: - Seunghwa Kang (https://github.com/seunghwak) - Brad Rees (https://github.com/BradReesWork) - Jordan Jacobelli (https://github.com/Ethyling) - Ray Douglass (https://github.com/raydouglass) URL: #1510
…rtex degrees (#1447) Partially addresses Issue #1442 Update graph primitives used by PageRank, Katz Centrality, BFS, and SSSP to launch 3 different kernels based on vertex degrees to address thread divergence issue. In addition, cut memory footprint of the VertexFrontier class used by BFS & SSSP. The following highlights performance improvement with this optimization. R-mat 2^25 vertices 2^25 * 32 edges PageRank: 7.66, 7.42, 8.83, 8.83 seconds (the first two unweighted, the last two weighted, first & third without personalization)=> 1.07, 1.08, 1.36, 1.39 seconds Katz: 1.08, 1.94 seconds (unweighted, weighted)=> 0.243, 0.275 BFS: 1.32 seconds=> 0.251 R-mat 2^25 vertices 2^25 * 16 edges SSSP: 1.89 seconds (memory allocation fails with the edge factor of 32)=> 0.317 And now SSSP also works with 2^25 vertices 2^25 * 32 edges with the memory footprint improvement and it took 0.514 sec. Still needs additional optimizations to reach the target performance 1. add BFS & SSSP specific optimizations (the current implementation assumes general reduction operations while BFS can pick any source vertex if a vertex is discovered by multiple source vertices and SSSP picks the one with the minimum edge weight, these pure function reduction operations allow additional optimizations). 2. Launch 3 different kernels in multiple streams to recover parallelism when the frontier size is relatively small (currently three kernels are queued in a single stream, and this leads to up to 3x decrease in parallelism) Authors: - Seunghwa Kang (https://github.com/seunghwak) Approvers: - Alex Fender (https://github.com/afender) - Chuck Hastings (https://github.com/ChuckHastings) - Brad Rees (https://github.com/BradReesWork) URL: #1447
…output (#1376) Solves: #1373 Authors: - Hugo Linsenmaier (https://github.com/hlinsen) Approvers: - Chuck Hastings (https://github.com/ChuckHastings) - Andrei Schaffer (https://github.com/aschaffer) - Alex Fender (https://github.com/afender) URL: #1376
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
aschaffer:branch-0.19 <- rapidsai:branch-0.19