| commit | 584b18eb4df61ccd447db2dfe8c8a7901f8c8598 | |
| author | Ondřej Bílka <neleai@seznam.cz> | |
| Sat, 14 Dec 2013 18:33:56 +0000 (14 19:33 +0100) | ||
| committer | Ondřej Bílka <neleai@seznam.cz> | |
| Sat, 14 Dec 2013 19:08:13 +0000 (14 20:08 +0100) | ||
| tree | 8240dbf408eadda74685f951e36f8885f77c2f77 | treesnapshot (tar.gz zip) |
| parent | 8a5c7897dd1c52ca74b06aaf5a3bacf0919c97aa | commitdiff |
Add strstr with unaligned loads. Fixes bug 12100.
A sse42 version of strstr used pcmpistr instruction which is quite
ineffective. A faster way is look for pairs of characters which is uses
sse2, is faster than pcmpistr and for real strings a pairs we look for
are relatively rare.
For linear time complexity we use buy or rent technique which switches
to two-way algorithm when superlinear behaviour is detected.
A sse42 version of strstr used pcmpistr instruction which is quite
ineffective. A faster way is look for pairs of characters which is uses
sse2, is faster than pcmpistr and for real strings a pairs we look for
are relatively rare.
For linear time complexity we use buy or rent technique which switches
to two-way algorithm when superlinear behaviour is detected.