Using bit masks for high-performance calculations

added by Szymon Warda
12/13/2016 11:08:04 AM

2 Kicks, 249 Views

Last time I've shown how I've gone from 34 hours to 11. This time we go faster. To go faster I have to do less. The current implementation of Similarity iterates over one vector and checks if that ingredient exists in the second one. Since those vectors are sparse the chance of a miss is big.