<br><br><div class="gmail_quote">2011/5/22 Reid Kleckner <span dir="ltr"><<a href="mailto:reid.kleckner@gmail.com">reid.kleckner@gmail.com</a>></span><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im">When asking this type of question, you should be specific about how</div>
you built the program, ie did you use clang, llvm-gcc, or dragonegg,<br>
and which options did you use. From your message, I can't tell if you<br>
built at O0 or O3.<br>
<br>
In this case, no, LLVM does not have any auto-vectorization<br>
optimizations. However, LLVM does have good support for vector<br>
intrinsics, so if you use xmmintrin.h you should be able to get good<br>
performance.<br>
<font color="#888888"><br>
Reid<br></font></blockquote><div> </div></div>Thanks, for your explanation. It is very useful for me.<div><br></div><div>--</div><div>Serg Anohovsky. </div>