在单节点设备上使用rmm

import rmm
from sklearn.datasets import make_classification

import xgboost as xgb

# Initialize RMM pool allocator
rmm.reinitialize(pool_allocator=True)
# Optionally force XGBoost to use RMM for all GPU memory allocation, see ./README.md
# xgb.set_config(use_rmm=True)

X, y = make_classification(n_samples=10000, n_informative=5, n_classes=3)
dtrain = xgb.DMatrix(X, label=y)

params = {
    "max_depth": 8,
    "eta": 0.01,
    "objective": "multi:softprob",
    "num_class": 3,
    "tree_method": "hist",
    "device": "cuda",
}
# XGBoost will automatically use the RMM pool allocator
bst = xgb.train(params, dtrain, num_boost_round=100, evals=[(dtrain, "train")])