Re: gcc4 optimization problems on intel mac
Re: gcc4 optimization problems on intel mac
- Subject: Re: gcc4 optimization problems on intel mac
- From: Nik Reiman <email@hidden>
- Date: Wed, 24 Jan 2007 21:25:49 +0100
On Jan 24, 2007, at 3:43 AM, Jim Wintermyre wrote:
Under *any* setting except -O0, these strange behaviors would
reappear, though without optimization everything runs normally.
Other optimization- related settings, such as loop unrolling, do
not seem to effect the software's stability.
I recall an issue I had (in an AU plugin), where another
optimization setting did seem to interact with the optimization
level. Specifically, I was having the problem only when the
optimization level was greater than -O0, AND "Enforce Strict
Aliasing" was enabled. In this case it wasn't an Intel-specific
issue, but the point is sometimes the settings do seem to affect
each other subtly.
Strict aliasing is tricky -- it's a nice performance optimization in
some cases, but the catch is that it's unsafe for certain code and
gcc can't warn you about all of the cases in which it's unsafe. In
other words, if you use it you can easily run into problems like this
where something goes wrong and you have to do a lot of debugging on a
release build to track it down. If that's something you'd rather
avoid (I know I would), it's probably best to avoid using strict
aliasing.
-Eric
FWIW, "enforce strict aliasing" is disabled for this project.
Anyways, thanks all for your suggestions... I'll try enabling more
compiler warnings and hacking through some of the code to see if
perhaps some uninitialized pointer might be causing trouble for us.
Cheers,
Nik
--
Nik Reiman :: email@hidden
Teragon Audio :: http://www.teragon.org
_______________________________________________
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