why dgl-ke so much faster than openke in test set

[finalTestin]

I do this in one GPU dglke_train --model_name TransE_l1 --dataset wn18 --batch_size 1000 --neg_sample_size 20 --hidden_dim 10000 --gamma 19.9 --lr 0.25 --max_step 100 --log_interval 100 --batch_size_eval 16 -adv --regularization_coef 1.00E-09 --test --num_thread 1 --num_proc 1 --gpu 0 in testing 100 times faster than openke，how dgl-ke did this

[finalTestin]

I do this in one GPU dglke_train --model_name TransE_l1 --dataset wn18 --batch_size 1000 --neg_sample_size 20 --hidden_dim 10000 --gamma 19.9 --lr 0.25 --max_step 100 --log_interval 100 --batch_size_eval 16 -adv --regularization_coef 1.00E-09 --test --num_thread 1 --num_proc 1 --gpu 0 in testing 100 times faster than openke，how dgl-ke did this

[classicsong]

You can take a look of our paper.