commit | b8ac5ac91c8be88023d1b9995630ef02216b40d4 | [log] [tgz] |
---|---|---|
author | philippe <philippe@a5019735-40e9-0310-863c-91ae7b9d1cf9> | Tue Apr 22 22:07:01 2014 +0000 |
committer | philippe <philippe@a5019735-40e9-0310-863c-91ae7b9d1cf9> | Tue Apr 22 22:07:01 2014 +0000 |
tree | c39eef7ce4366a36abf28635043f7d6dd789bb9d | |
parent | adfff76735ea5bca870dec8f382b225728e1b32d [diff] |
* Implement optimisation in storev8/16/64 similar to optimisations in storev32 See comments in MC_(helperc_STOREV8) for a detailed explanation. On modern cpu, this gives a few percent speed improvement for some perf tests, and reduces the nr of undistinguished secondary map. git-svn-id: svn://svn.valgrind.org/valgrind/trunk@13904 a5019735-40e9-0310-863c-91ae7b9d1cf9