diff options
| author | Arseny Kapoulkine <arseny.kapoulkine@gmail.com> | 2014-08-26 06:10:52 +0000 |
|---|---|---|
| committer | Arseny Kapoulkine <arseny.kapoulkine@gmail.com> | 2014-08-26 06:10:52 +0000 |
| commit | 211212c9861834c2f8327a1efe6690a893365823 (patch) | |
| tree | b33e1255d1f593a4ed82299bcb133bcfe90346ef /docs/samples/weekly-shift_jis.xml | |
| parent | 0f5101bfc5355cbc0c5525403be0073bc614b390 (diff) | |
Unroll performance-critical loops 4x
Use a special macro that unrolls the loop body and uses static branch prediction
to improve code generation.
This increases performance across all data sets from benchmark; clang x64 is 10%-40%
faster, clang x86 is 5%-20% faster, msvc is 5%-10% faster.
git-svn-id: https://pugixml.googlecode.com/svn/trunk@1008 99668b35-9821-0410-8761-19e4c4f06640
Diffstat (limited to 'docs/samples/weekly-shift_jis.xml')
0 files changed, 0 insertions, 0 deletions