Why sklearn's kNN classifer runs so fast while the number of my training samples and test samples are large

Question

According to my understanding, for each test sample, kNN classifier algorithm computes the distances between the current test sample and all the training samples and select a ce