numpy/unpackbits-popcount
| Rule | unpackbits-popcount |
|---|---|
| Pack | NumPy |
| Severity | warning |
| Fix | none, edit by hand |
| Library versions | NumPy 2.0 and later |
Unpacking every byte into eight bits to count the set ones was 37 times slower. Use np.bitwise_count.
ones = np.unpackbits(masks.view(np.uint8)).sum()ones = np.bitwise_count(masks).sum()What the rule reports
Section titled “What the rule reports”The rule reports np.unpackbits(...).sum(), np.sum(np.unpackbits(...)) and np.count_nonzero(np.unpackbits(...)). np.bitwise_count, added in NumPy 2.0, counts the set bits of each value, and ran 37 times faster on 1,000,000 uint64 values.
How to fix it
Section titled “How to fix it”xpd has no automatic fix. Because the view, reshape and axis around unpackbits vary. Make the change by hand.
Message
Section titled “Message”xpd prints this message with the issue. A name that starts with $ stands for code from your file.
This unpacks every byte into eight before it counts the set bits. Write $NP.bitwise_count(x), summed as needed, which ran 37 times faster on 1,000,000 uint64 values.Browse the other rules of the NumPy pack.