RandomForestClassifier got error when training on relative large dataset

[justinuliu]

(accelweka) ~ $ weka -main weka.Run .RandomRBF -n 5000 -a 5000 > RBFa5kn5k.arff
(accelweka) ~ $ weka -memory 32g -main weka.Run weka.classifiers.rapids.CuMLDaskClassifier -learner RandomForestClassifier -t $(pwd)/RBFa5kn5k.arff
java.lang.Exception: Traceback (most recent call last):
File "/home/justinliu/wekafiles/packages/wekaRAPIDS/resources/py/pyRapidsServer.py", line 320, in execute_script
exec(script, _global_env)
File "", line 7, in
File "/home/justinliu/miniconda3/envs/accelweka/lib/python3.9/site-packages/cuml/dask/ensemble/randomforestclassifier.py", line 263, in fit
self._fit(model=self.rfs,
File "/home/justinliu/miniconda3/envs/accelweka/lib/python3.9/site-packages/cuml/dask/ensemble/base.py", line 158, in _fit
wait_and_raise_from_futures(futures)
File "/home/justinliu/miniconda3/envs/accelweka/lib/python3.9/site-packages/cuml/dask/common/utils.py", line 162, in wait_and_raise_from_futures
raise_exception_from_futures(futures)
File "/home/justinliu/miniconda3/envs/accelweka/lib/python3.9/site-packages/cuml/dask/common/utils.py", line 151, in raise_exception_from_futures
raise RuntimeError("%d of %d worker jobs failed: %s" % (
RuntimeError: 1 of 1 worker jobs failed: radix_sort: failed on 2nd step: cudaErrorInvalidValue: invalid argument

    at weka.classifiers.rapids.CuMLDaskClassifier.buildClassifier(CuMLDaskClassifier.java:864)
    at weka.classifiers.evaluation.Evaluation.evaluateModel(Evaluation.java:1632)
    at weka.classifiers.Evaluation.evaluateModel(Evaluation.java:668)
    at weka.classifiers.AbstractClassifier.runClassifier(AbstractClassifier.java:141)
    at weka.classifiers.AbstractClassifier.run(AbstractClassifier.java:547)
    at weka.Run.main(Run.java:349)

This error occurs on GTX1080Ti but not on GTX1660 SUPER