Re: GCC 4 and strict aliasing
Re: GCC 4 and strict aliasing
- Subject: Re: GCC 4 and strict aliasing
- From: Eric Christopher <email@hidden>
- Date: Sat, 18 Mar 2006 02:40:08 -0800
Awesome, thanks! I will file the bug reports as you suggested. I've
also noticed that Xcode 2.2 / GCC 4 generates rather sub-optimal
code for Intel, with a lot of unnecessary extra register
movements. Is this a known problem?
Some of them are. If you like, go ahead and try to reduce a testcase
to show anything particularly bad.
2) Apple's GCC should not turn on -fstrict-aliasing when -ftree-
vectorize -fno-strict-aliasing is passed.
Don't file this one - the vectorizer needs the information to
vectorize properly (a good reason to make your code aliasing safe :)
3) The GCC man page should document that -fstrict-aliasing is
turned off by default on Mac OS X and that -ftree-vectorize turns
it on.
Don't worry about filing this one. It's fixed already.
thanks.
-eric
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden