Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
7,046 Commits 32 Branches 59 Tags
9b88495840785d5f8bd51985139592bae50d2ce7
Commit Graph
100 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Jiaming Yuan 9b88495840 [multi] Implement weight feature importance. (#10700) 2024-08-22 02:06:47 +08:00
Jiaming Yuan 402e7837fb Fix potential race in feature constraint. (#10719) 2024-08-21 16:50:31 +08:00
Jiaming Yuan 508ac13243 Check cub errors. (#10721) 
- Make sure cuda error returned by cub scan is caught.
- Avoid temporary buffer allocation in thrust device vector.
 2024-08-21 02:50:26 +08:00
Jiaming Yuan fd365c147e [doc] Brief note about RMM SAM allocator. [skip ci] (#10712) 2024-08-17 04:21:39 +08:00
Jiaming Yuan ec3f327c20 Add managed memory allocator. (#10711) 2024-08-17 03:02:34 +08:00
Jiaming Yuan 8d7fe262d9 [EM] Enable access to the number of batches. (#10691) 
- Expose `NumBatches` in `DMatrix`.
- Small cleanup for removing legacy CUDA stream and ~force CUDA context initialization~.
- Purge old external memory data generation code.
 2024-08-17 02:59:45 +08:00
Jiaming Yuan 033a666900 [EM] Log the page size of ellpack. (#10713) 2024-08-17 01:35:47 +08:00
Jiaming Yuan abe65e3769 Reduce thread contention in column split histogram test. (#10708) 2024-08-17 01:00:32 +08:00
Jiaming Yuan 2258bc870d Add more tests and doc for QDM. (#10692) 2024-08-16 23:30:04 +08:00
Jiaming Yuan 582ea104b5 [EM] Enable prediction cache for GPU. (#10707) 
- Use `UpdatePosition` for all nodes and skip `FinalizePosition` when external memory is used.
- Create `encode/decode` for node position, this is just as a refactor.
- Reuse code between update position and finalization.
 2024-08-15 21:41:59 +08:00
Jiaming Yuan 2ecc85ffad [EM] Support ExtMemQdm in the GPU predictor. (#10694) 2024-08-13 12:21:11 +08:00
Jiaming Yuan 43704549a2 [coll] Reduce the amount of open files (socket). (#10693) 
Reduce the chance of hitting `Failed to call `socket`: Too many open files`.
 2024-08-13 05:23:49 +08:00
Jiaming Yuan d414fdf2e7 [EM] Add GPU version of the external memory QDM. (#10689) 2024-08-10 10:49:43 +08:00
Jiaming Yuan 7bccc1ea2c [EM] CPU implementation for external memory QDM. (#10682) 
- A new DMatrix type.
- Extract common code into a new QDM base class.

Not yet working:
- Not exposed to the interface yet, will wait for the GPU implementation.
- ~No meta info yet, still working on the source.~
- Exporting data to CSR is not supported yet.
 2024-08-09 09:38:02 +08:00
Jiaming Yuan cc3b56fc37 Cleanup GPU Hist tests. (#10677) 
* Cleanup GPU Hist tests.

- Remove GPU Hist gradient sampling test. The same properties are tested in the gradient
  sampler test suite.
- Move basic histogram tests into the histogram test suite.
- Remove the header inclusion of the `updater_gpu_hist.cu` in tests.
 2024-08-06 11:50:44 +08:00
Jiaming Yuan 6ccf116601 [dask] Reduce the flakiness of tests. (#10678) 2024-08-06 06:04:10 +08:00
Jiaming Yuan 3d8107adb8 Support doc link for the sklearn module. (#10287) 2024-08-06 02:35:32 +08:00
Jiaming Yuan a269055b2b [coll] Use loky for tests. (#10676) 
This makes the tests easier to run and debug. In addition, they can now work on Windows as
well.
 2024-08-03 07:33:42 +08:00
Jiaming Yuan a185b693dc Reduce warnings and flakiness in tests. (#10659) 
- Fix warnings in tests.
- Try to reduce the flakiness of dask test.
 2024-08-03 07:32:47 +08:00
Jiaming Yuan 2e7ba900ef [CI] Add timeout limit to JVM tests. (#10673) 2024-08-03 01:51:13 +08:00
Jiaming Yuan 574c20dc1d Enable CI build for the federated-secure branch. (#10671) 2024-08-02 22:13:17 +08:00
Jiaming YuanandPhilip Hyunsu Cho 77c844cef7 Reduce thread contention in column split tests. (#10658) 
---------

Co-authored-by: Philip Hyunsu Cho <chohyu01@cs.washington.edu>
 2024-08-01 18:36:46 +08:00
Jiaming YuanandHyunsu Cho fb77ed7603 [CI] Fix Python wheel workflow. (#10649) 
* [CI] Fix Python wheel workflow.

* Use Python 3.10 for building wheels

---------

Co-authored-by: Hyunsu Cho <phcho@nvidia.com>
 2024-07-30 10:13:47 -07:00
Jiaming Yuan 827d0e8edb [breaking] Bump Python requirement to 3.10. (#10434) 
- Bump the Python requirement.
- Fix type hints.
- Use loky to avoid deadlock.
- Workaround cupy-numpy compatibility issue on Windows caused by the `safe` casting rule.
- Simplify the repartitioning logic to avoid dask errors.
 2024-07-30 17:31:06 +08:00
Jiaming Yuan 449be7a402 Quick fix for clang-tidy error. (#10641) 2024-07-26 18:21:16 +08:00
Jiaming Yuan fcae6301ec [dask] Disable broadcast in the scatter call. (#10632) 2024-07-25 04:16:34 +08:00
Jiaming Yuan 485d90218c Catch exceptions during file read. (#10623) 2024-07-23 03:48:19 +08:00
Jiaming Yuan a19bbc9be5 Avoid caching allocator for large allocations. (#10582) 2024-07-23 03:48:03 +08:00
Jiaming Yuan b2cae34a8e Fix integer overflow. (#10615) 2024-07-23 02:13:15 +08:00
Jiaming Yuan 6d9fcb771e Move device histogram storage into histogram.cuh. (#10608) 2024-07-21 14:10:13 +08:00
Jiaming Yuan cb62f9e73b [EM] Prevent init with CUDA malloc resource. (#10606) 2024-07-21 05:08:29 +08:00
Jiaming Yuan 0846ad860c Optionally skip cupy on windows. (#10611) 2024-07-20 22:12:12 +08:00
Jiaming Yuan 344ddeb9ca Drop support for CUDA legacy stream. (#10607) 2024-07-20 06:14:56 +08:00
Jiaming Yuan 7ab93f3ce3 [CI] Fix test environment. (#10609) 
* [CI] Fix test environment.

* Remove shell.

* Remove.

* Update Dockerfile.i386
 2024-07-18 10:04:17 -07:00
Jiaming Yuan 292bb677e5 [EM] Support mmap backed ellpack. (#10602) 
- Support resource view in ellpack.
- Define the CUDA version of MMAP resource.
- Define the CUDA version of malloc resource.
- Refactor cuda runtime API wrappers, and add memory access related wrappers.
- gather windows macros into a single header.
 2024-07-18 08:20:21 +08:00
Jiaming Yuan e9fbce9791 Refactor DeviceUVector. (#10595) 
Create a wrapper instead of using inheritance to avoid inconsistent interface of the class.
 2024-07-18 03:33:01 +08:00
Jiaming Yuan a6a8a55ffa Merge approx tests. (#10583) 2024-07-16 19:03:48 +08:00
Jiaming Yuan 5a92ffe3ca Partial fix for CTK 12.5 (#10574) 2024-07-16 17:41:50 +08:00
Jiaming Yuan bbd308595a [jvm-packages] Bump rapids version. (#10588) 2024-07-15 20:21:25 +08:00
Jiaming Yuan 5fea9d24f2 Small cleanup for CMake scripts. (#10573) 
- Remove rabit.
 2024-07-12 05:18:23 +08:00
Jiaming Yuan 6c403187ec Fix column split race condition. (#10572) 2024-07-12 01:07:12 +08:00
Jiaming Yuan 1ca4bfd20e Avoid thrust vector initialization. (#10544) 
* Avoid thrust vector initialization.

- Add a wrapper for rmm device uvector.
- Split up the `Resize` method for HDV.
 2024-07-11 17:29:27 +08:00
Jiaming Yuan 89da9f9741 [fed] Split up federated test CMake file. (#10566) 
- Collect all federated test files into the same directory.
- Independently list the files.
 2024-07-11 13:09:18 +08:00
Jiaming Yuan 5f910cd4ff [EM] Handle base idx in GPU histogram. (#10549) 2024-07-11 03:26:30 +08:00
Jiaming Yuan 34b154c284 Avoid the use of size_t in the partitioner. (#10541) 
- Avoid the use of size_t in the partitioner.
- Use `Span` instead of `Elem` where `node_id` is not needed.
- Remove the `const_cast`.
- Make sure the constness is not removed in the `Elem` by making it reference only.

size_t is implementation-defined, which causes issue when we want to pass pointer or span.
 2024-07-11 00:43:08 +08:00
Jiaming Yuan baba3e9eb0 Fix empty partition. (#10559) 2024-07-10 13:01:47 +08:00
Jiaming Yuan 8e2b874b4c [doc] Add notes about RMM and device ordinal. [skip ci] (#10562) 
- Remove the experimental tag, we have been running it for a long time now.
- Add notes about avoiding set CUDA device.
- Add link in parameter.
 2024-07-10 13:00:57 +08:00
Jiaming Yuan 3ec74a1ba9 [doc] Add build_info to autodoc. [skip ci] (#10551) 2024-07-10 04:05:20 +08:00
Jiaming Yuan 2266db17d1 [R] Update roxygen. (#10556) 2024-07-08 17:02:46 +08:00
Jiaming Yuan 00264eb72b [EM] Basic distributed test for external memory. (#10492) 2024-07-06 01:15:20 +08:00
Jiaming Yuan 620b2b155a Cache GPU histogram kernel configuration. (#10538) 2024-07-04 15:38:59 +08:00
Jiaming Yuan cd1d108c7d [doc] Fix learning to rank tutorial. [skip ci] (#10539) 2024-07-03 22:52:26 +08:00
Jiaming Yuan 6243e7c43d [doc] Update link to release notes. [skip ci] (#10533) 2024-07-03 12:16:53 +08:00
Jiaming Yuan 628411a654 Enhance the threadpool implementation. (#10531) 
- Accept an initialization function.
- Support void return tasks.
 2024-07-03 12:13:27 +08:00
Jiaming Yuan 9cb4c938da [EM] Move prefetch in reset into the end of the iteration. (#10529) 2024-07-03 03:48:18 +08:00
Jiaming Yuan e537b0969f Fix boolean array for arrow-backed DF. (#10527) 2024-07-02 17:02:54 +08:00
Jiaming Yuan d33043a348 [coll] Allow using local host for testing. (#10526) 
- Don't try to retrieve the IP address if a host is specified.
- Fix compiler deprecation warning.
 2024-07-02 15:34:38 +08:00
Jiaming Yuan a39fef2c67 [fed] Fixes for the encrypted GRPC backend. (#10503) 2024-07-02 15:15:12 +08:00
Jiaming Yuan 5f0c1e902b Small cleanup for error message. (#10502) 
- The `Fail` function can handle file location automatically.
- Report concatenated error for connection poll.
- Typos.
 2024-07-02 13:36:41 +08:00
Jiaming Yuan e8a962575a [EM] Allow staging ellpack on host for GPU external memory. (#10488) 
- New parameter `on_host`.
- Abstract format creation and stream creation into policy classes.
 2024-06-28 04:42:18 +08:00
Jiaming Yuan 824fba783e Remove support for deprecated format in Python. (#10490) 2024-06-27 11:31:53 +08:00
Jiaming Yuan 2d88d17008 Remove deprecated DeviceQuantileDMatrix. (#10491) 2024-06-27 11:30:51 +08:00
Jiaming Yuan 26eb68859f Consistently report error in tests. (#10453) 2024-06-21 14:35:22 +08:00
Jiaming Yuan b38c7fe2ce 2.1.0 release news.(#10378) 2024-06-21 14:34:53 +08:00
Jiaming Yuan 2b400b18d5 Small cleanup for rowset collection. (#10401) 2024-06-19 18:06:23 +08:00
Jiaming Yuan e5f1720656 [EM] Avoid writing cut matrix to cache. (#10444) 2024-06-19 18:03:38 +08:00
Jiaming Yuan 63418d2f35 Link CMAKE_DL_LIBS when dlopen is used. (#10447) 2024-06-19 15:06:58 +08:00
Jiaming Yuanhcho3Philip Hyunsu Cho
 		b9e5229ff2 Update rapids (#10435) 
* [CI] Update RAPIDS to latest stable

* RMM.

---------

Co-authored-by: hcho3 <2532981+hcho3@users.noreply.github.com>
Co-authored-by: Philip Hyunsu Cho <chohyu01@cs.washington.edu>
 2024-06-18 05:01:57 +08:00
Jiaming Yuan b4cc350ec5 Fix categorical data with external memory. (#10433) 2024-06-18 04:34:54 +08:00
Jiaming Yuan a8ddbac163 [doc] Fixes for external memory document. (#10426) 2024-06-18 03:10:49 +08:00
Jiaming YuanandPhilip Hyunsu Cho 320e7c2041 [CI] Enable CI binary build for the vertical federated learning branch. (#10417) 
Co-authored-by: Philip Hyunsu Cho <chohyu01@cs.washington.edu>
 2024-06-17 17:01:04 +08:00
Jiaming YuanandPhilip Hyunsu Cho 6c83c8c2ef Allow blocking launch of federated tracker. (#10414) 
---------

Co-authored-by: Philip Hyunsu Cho <chohyu01@cs.washington.edu>
 2024-06-16 01:43:53 +08:00
Jiaming YuanandPhilip Hyunsu Cho 49e25cfb36 Allow unaligned pointer if the array is empty. (#10418) 
Co-authored-by: Philip Hyunsu Cho <chohyu01@cs.washington.edu>
 2024-06-15 19:10:21 +08:00
Jiaming YuanandPhilip Hyunsu Cho bbff74d2ff [dask] Workaround the tokenizer by changing the scatter function. (#10419) 
---------

Co-authored-by: Philip Hyunsu Cho <chohyu01@cs.washington.edu>
 2024-06-15 19:10:00 +08:00
Jiaming Yuan c9f5fcaf21 [col] Small cleanup to federated comm. (#10397) 2024-06-07 21:19:04 +08:00
Jiaming Yuan 9f6608d6aa Add python 3.12 classifier. (#10381) 2024-06-04 18:02:59 +08:00
Jiaming Yuan 43a57c4a85 Bump development version to 2.2. (#10376) 2024-06-04 12:59:16 +08:00
Jiaming Yuan 979e392deb Fix warnings in GPU dask tests. (#10358) 2024-06-04 12:58:58 +08:00
Jiaming Yuan 0808e50ae8 Sync stream in ellpack format. (#10374) 2024-06-04 12:58:26 +08:00
Jiaming Yuan 7f3e92d71a Bump the cache github action to 4.0.2. (#10377) 2024-06-03 16:52:21 +08:00
Jiaming Yuan 4f48647932 Fix typo. (#10353) 2024-06-02 02:07:55 +08:00
Jiaming Yuan 92cba25fe2 Remove reference to R win64 MSVC build. (#10355) 2024-06-02 02:06:35 +08:00
Jiaming Yuan d2d01d977a Remove unnecessary fetch operations in external memory. (#10342) 2024-05-31 13:16:40 +08:00
Jiaming Yuan c2e3d4f3cd [dask] Update dask demo for using the new dask backend. (#10347) 2024-05-31 08:03:20 +08:00
Jiaming Yuan e6eefea5e2 [coll] Move the rabit poll helper. (#10349) 2024-05-31 08:02:21 +08:00
Jiaming Yuan 2de67f0050 [coll] Prevent race during error check. (#10319) 2024-05-28 15:43:16 -07:00
Jiaming YuanandPhilip Hyunsu Cho 7354955cbb Test federated plugin using GitHub action. (#10336) 
Co-authored-by: Philip Hyunsu Cho <chohyu01@cs.washington.edu>
 2024-05-29 02:28:14 +08:00
Jiaming Yuan 5627af6b21 [coll] Increase timeout limit. (#10332) 2024-05-28 10:20:49 +08:00
Jiaming Yuan 15eb553c1f [doc] Add a coarse map for XGBoost features to assist development. [skip ci] (#10310) 2024-05-23 14:25:15 +08:00
Jiaming Yuan 966dc81788 [coll] Keep the tracker alive during initialization error. (#10306) 2024-05-23 11:13:59 +08:00
Jiaming Yuan d5fcbee44b Add timeout for distributed tests. (#10315) 2024-05-23 11:11:49 +08:00
Jiaming Yuan 1b25d23583 [JVM-packages] Prevent memory leak. (#10307) 2024-05-22 13:47:59 +08:00
Jiaming Yuan 7a54ca41c9 [CI] Bump checkout action version. (#10305) 2024-05-21 16:38:20 +08:00
Jiaming Yuan a5a58102e5 Revamp the rabit implementation. (#10112) 
This PR replaces the original RABIT implementation with a new one, which has already been partially merged into XGBoost. The new one features:
- Federated learning for both CPU and GPU.
- NCCL.
- More data types.
- A unified interface for all the underlying implementations.
- Improved timeout handling for both tracker and workers.
- Exhausted tests with metrics (fixed a couple of bugs along the way).
- A reusable tracker for Python and JVM packages.
 2024-05-20 11:56:23 +08:00
Jiaming Yuan ba9b4cb1ee Fix pylint. (#10296) 2024-05-17 13:28:39 +08:00
Jiaming Yuan 835e59e538 Use a thread pool for external memory. (#10288) 2024-05-16 19:32:12 +08:00
Jiaming Yuan ca1d04bcb7 Release data in cache. (#10286) 2024-05-14 14:20:19 +08:00
Jiaming Yuan f1f69ff10e [CI] Fixes for using the latest modin. (#10285) 2024-05-14 12:13:35 +08:00
Jiaming Yuan 871fabeee3 [doc][dask] Update notes about k8s. (#10271) 2024-05-14 04:21:02 +08:00
Jiaming YuanandPhilip Hyunsu Cho d81e319e78 Fixes for the latest pandas. (#10266) 
Co-authored-by: Philip Hyunsu Cho <chohyu01@cs.washington.edu>
 2024-05-12 11:15:46 +08:00