Rory Mitchell 71d3b2e036

Fuse gpu_hist all-reduce calls where possible (#7867)

2022-05-17 13:27:50 +02:00

..

evaluate_splits.cu

Small refactor to categoricals (#7858)

2022-05-05 17:47:02 +02:00

evaluate_splits.cuh

Fuse gpu_hist all-reduce calls where possible (#7867)

2022-05-17 13:27:50 +02:00

evaluator.cu

Small refactor to categoricals (#7858)

2022-05-05 17:47:02 +02:00

expand_entry.cuh

Support optimal partitioning for GPU hist. (#7652)

2022-02-15 03:03:12 +08:00

feature_groups.cu

Split Features into Groups to Compute Histograms in Shared Memory (#5795)

2020-07-07 15:04:35 +12:00

feature_groups.cuh

Split Features into Groups to Compute Histograms in Shared Memory (#5795)

2020-07-07 15:04:35 +12:00

gradient_based_sampler.cu

Fix external memory with gpu_hist and subsampling combination bug. (#7481)

2021-12-24 11:15:35 +08:00

gradient_based_sampler.cuh

Fix external memory with gpu_hist and subsampling combination bug. (#7481)

2021-12-24 11:15:35 +08:00

histogram.cu

Fuse gpu_hist all-reduce calls where possible (#7867)

2022-05-17 13:27:50 +02:00

histogram.cuh

[breaking] Use integer atomic for GPU histogram. (#7180)

2021-08-28 05:17:05 +08:00

row_partitioner.cu

Enable compiling with system cub. (#7232)

2021-09-17 14:28:18 +08:00

row_partitioner.cuh

[POC] Experimental support for l1 error. (#7812)

2022-04-26 21:41:55 +08:00