<div dir="ltr">This sounds very similar to the immintrin.h etc headers that Intel provides, which IMO are great. We don't have to implement everything as a builtin in the compiler. We can insulate ourselves with a header layer.<div>
<br></div><div>GCC's __builtins aren't standard, so having a real document seems like a step forward.</div><div><br></div><div>-- from a peanut gallery somewhere</div></div><div class="gmail_extra"><br><br><div class="gmail_quote">
On Wed, Jun 25, 2014 at 2:44 PM, Saleem Abdulrasool <span dir="ltr"><<a href="mailto:compnerd@compnerd.org" target="_blank">compnerd@compnerd.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Im sure that Im missing some context here. I am not sure what using this wrapper header to map the intrinsics to a __builtin_ form really gains. It adds an additional header that needs to be maintained, and requires that users include this header for the intrinsics.<br>
<br>
At the very least, similar approaches to this on the Windows side have proven to be less than ideal.<br>
<br>
Having some additional context around why this is preferable would be nice.<br>
<br>
<a href="http://reviews.llvm.org/D4296" target="_blank">http://reviews.llvm.org/D4296</a><br>
<br>
<br>
<br>
_______________________________________________<br>
cfe-commits mailing list<br>
<a href="mailto:cfe-commits@cs.uiuc.edu">cfe-commits@cs.uiuc.edu</a><br>
<a href="http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits" target="_blank">http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits</a><br>
</blockquote></div><br></div>