Counting unique elements faster in Google BigQuery with HyperLogLog

With HyperLogLog algorithm, we trade off some accuracy to get the job done faster. But how much faster and at the expense of what level of accuracy?